WO1999001843A1 - Method for remote sensing analysis by decorrelation statistical analysis and hardware therefor - Google Patents
Method for remote sensing analysis by decorrelation statistical analysis and hardware therefor Download PDFInfo
- Publication number
- WO1999001843A1 WO1999001843A1 PCT/US1998/013524 US9813524W WO9901843A1 WO 1999001843 A1 WO1999001843 A1 WO 1999001843A1 US 9813524 W US9813524 W US 9813524W WO 9901843 A1 WO9901843 A1 WO 9901843A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- scenes
- specfral
- classification
- remote
- decorrelated
- Prior art date
Links
- 238000000034 method Methods 0.000 title claims abstract description 162
- 238000007619 statistical method Methods 0.000 title claims abstract description 65
- 238000004458 analytical method Methods 0.000 title claims description 42
- 230000003595 spectral effect Effects 0.000 claims abstract description 315
- 238000001228 spectrum Methods 0.000 claims abstract description 117
- 238000000513 principal component analysis Methods 0.000 claims abstract description 53
- 238000002360 preparation method Methods 0.000 claims abstract description 21
- 239000013598 vector Substances 0.000 claims description 80
- 230000003287 optical effect Effects 0.000 claims description 57
- 238000005259 measurement Methods 0.000 claims description 51
- 230000005855 radiation Effects 0.000 claims description 48
- 238000012935 Averaging Methods 0.000 claims description 19
- 230000003278 mimic effect Effects 0.000 claims description 18
- 238000004364 calculation method Methods 0.000 claims description 9
- 238000000354 decomposition reaction Methods 0.000 claims description 8
- 230000005670 electromagnetic radiation Effects 0.000 claims description 8
- 239000006185 dispersion Substances 0.000 claims description 6
- 238000004519 manufacturing process Methods 0.000 claims description 4
- 210000000349 chromosome Anatomy 0.000 description 83
- 239000000463 material Substances 0.000 description 41
- 238000003384 imaging method Methods 0.000 description 33
- 239000011159 matrix material Substances 0.000 description 23
- 238000010521 absorption reaction Methods 0.000 description 22
- 230000009102 absorption Effects 0.000 description 21
- 239000000523 sample Substances 0.000 description 17
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Chemical compound O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 16
- 239000002689 soil Substances 0.000 description 15
- 230000005284 excitation Effects 0.000 description 14
- 238000004020 luminiscence type Methods 0.000 description 14
- 230000006870 function Effects 0.000 description 12
- 238000013459 approach Methods 0.000 description 11
- 239000000839 emulsion Substances 0.000 description 11
- 230000035945 sensitivity Effects 0.000 description 11
- 230000002596 correlated effect Effects 0.000 description 10
- 238000001514 detection method Methods 0.000 description 10
- 210000001508 eye Anatomy 0.000 description 10
- 239000000203 mixture Substances 0.000 description 10
- 230000008685 targeting Effects 0.000 description 10
- 238000011156 evaluation Methods 0.000 description 9
- 238000005286 illumination Methods 0.000 description 9
- 239000003086 colorant Substances 0.000 description 8
- 230000009467 reduction Effects 0.000 description 8
- 239000000975 dye Substances 0.000 description 7
- 238000002474 experimental method Methods 0.000 description 7
- 239000004973 liquid crystal related substance Substances 0.000 description 7
- 239000006096 absorbing agent Substances 0.000 description 6
- 230000008859 change Effects 0.000 description 6
- 239000002131 composite material Substances 0.000 description 6
- 238000013480 data collection Methods 0.000 description 6
- 230000000694 effects Effects 0.000 description 6
- 210000003917 human chromosome Anatomy 0.000 description 6
- 230000035882 stress Effects 0.000 description 6
- 230000008901 benefit Effects 0.000 description 5
- 238000006243 chemical reaction Methods 0.000 description 5
- 238000010276 construction Methods 0.000 description 5
- 239000013078 crystal Substances 0.000 description 5
- 229910052500 inorganic mineral Inorganic materials 0.000 description 5
- 238000000386 microscopy Methods 0.000 description 5
- 239000011707 mineral Substances 0.000 description 5
- 235000010755 mineral Nutrition 0.000 description 5
- 239000011435 rock Substances 0.000 description 5
- 241000894007 species Species 0.000 description 5
- 230000005540 biological transmission Effects 0.000 description 4
- 230000000903 blocking effect Effects 0.000 description 4
- 239000011575 calcium Substances 0.000 description 4
- 238000000576 coating method Methods 0.000 description 4
- 230000000875 corresponding effect Effects 0.000 description 4
- 238000010252 digital analysis Methods 0.000 description 4
- 238000009826 distribution Methods 0.000 description 4
- 230000005684 electric field Effects 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 238000011065 in-situ storage Methods 0.000 description 4
- 238000013507 mapping Methods 0.000 description 4
- 238000012986 modification Methods 0.000 description 4
- 230000004048 modification Effects 0.000 description 4
- 239000003973 paint Substances 0.000 description 4
- 238000011160 research Methods 0.000 description 4
- 239000007787 solid Substances 0.000 description 4
- 239000000126 substance Substances 0.000 description 4
- 239000011800 void material Substances 0.000 description 4
- 241000196324 Embryophyta Species 0.000 description 3
- 210000005252 bulbus oculi Anatomy 0.000 description 3
- 238000012937 correction Methods 0.000 description 3
- 238000007405 data analysis Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 230000018109 developmental process Effects 0.000 description 3
- 238000000295 emission spectrum Methods 0.000 description 3
- 239000007850 fluorescent dye Substances 0.000 description 3
- 239000010440 gypsum Substances 0.000 description 3
- 229910052602 gypsum Inorganic materials 0.000 description 3
- 238000009396 hybridization Methods 0.000 description 3
- 230000003993 interaction Effects 0.000 description 3
- 239000007788 liquid Substances 0.000 description 3
- 238000007781 pre-processing Methods 0.000 description 3
- 230000004044 response Effects 0.000 description 3
- 238000010186 staining Methods 0.000 description 3
- 230000009466 transformation Effects 0.000 description 3
- 230000000007 visual effect Effects 0.000 description 3
- 239000002028 Biomass Substances 0.000 description 2
- 208000031404 Chromosome Aberrations Diseases 0.000 description 2
- 206010010774 Constipation Diseases 0.000 description 2
- 239000003298 DNA probe Substances 0.000 description 2
- 229910000661 Mercury cadmium telluride Inorganic materials 0.000 description 2
- 230000004075 alteration Effects 0.000 description 2
- 230000004888 barrier function Effects 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 210000004027 cell Anatomy 0.000 description 2
- 238000000701 chemical imaging Methods 0.000 description 2
- 231100000005 chromosome aberration Toxicity 0.000 description 2
- 239000004567 concrete Substances 0.000 description 2
- 230000003247 decreasing effect Effects 0.000 description 2
- 230000001419 dependent effect Effects 0.000 description 2
- 238000003745 diagnosis Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000005611 electricity Effects 0.000 description 2
- 230000005672 electromagnetic field Effects 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 238000000799 fluorescence microscopy Methods 0.000 description 2
- 239000003673 groundwater Substances 0.000 description 2
- 238000007901 in situ hybridization Methods 0.000 description 2
- WPYVAWXEWQSOGY-UHFFFAOYSA-N indium antimonide Chemical compound [Sb]#[In] WPYVAWXEWQSOGY-UHFFFAOYSA-N 0.000 description 2
- 230000010354 integration Effects 0.000 description 2
- 238000011545 laboratory measurement Methods 0.000 description 2
- 230000005389 magnetism Effects 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 230000007935 neutral effect Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- PYWVYCXTNDRMGF-UHFFFAOYSA-N rhodamine B Chemical compound [Cl-].C=12C=CC(=[N+](CC)CC)C=C2OC2=CC(N(CC)CC)=CC=C2C=1C1=CC=CC=C1C(O)=O PYWVYCXTNDRMGF-UHFFFAOYSA-N 0.000 description 2
- 230000001932 seasonal effect Effects 0.000 description 2
- 238000004611 spectroscopical analysis Methods 0.000 description 2
- 238000012546 transfer Methods 0.000 description 2
- 238000012800 visualization Methods 0.000 description 2
- 239000002699 waste material Substances 0.000 description 2
- 208000002874 Acne Vulgaris Diseases 0.000 description 1
- 241000894006 Bacteria Species 0.000 description 1
- 229910021532 Calcite Inorganic materials 0.000 description 1
- OYPRJOBELJOOCE-UHFFFAOYSA-N Calcium Chemical compound [Ca] OYPRJOBELJOOCE-UHFFFAOYSA-N 0.000 description 1
- 244000025254 Cannabis sativa Species 0.000 description 1
- 241000195493 Cryptophyta Species 0.000 description 1
- 238000004971 IR microspectroscopy Methods 0.000 description 1
- 206010061217 Infestation Diseases 0.000 description 1
- 244000073231 Larrea tridentata Species 0.000 description 1
- 235000019738 Limestone Nutrition 0.000 description 1
- 241000721696 Lymantria Species 0.000 description 1
- 241000428199 Mustelinae Species 0.000 description 1
- 238000012952 Resampling Methods 0.000 description 1
- 244000166071 Shorea robusta Species 0.000 description 1
- 235000015076 Shorea robusta Nutrition 0.000 description 1
- XUIMIQQOPSSXEZ-UHFFFAOYSA-N Silicon Chemical compound [Si] XUIMIQQOPSSXEZ-UHFFFAOYSA-N 0.000 description 1
- 240000008042 Zea mays Species 0.000 description 1
- 235000005824 Zea mays ssp. parviglumis Nutrition 0.000 description 1
- 235000002017 Zea mays subsp mays Nutrition 0.000 description 1
- 206010000496 acne Diseases 0.000 description 1
- 239000006098 acoustic absorber Substances 0.000 description 1
- 230000003698 anagen phase Effects 0.000 description 1
- 239000010426 asphalt Substances 0.000 description 1
- 238000003556 assay Methods 0.000 description 1
- 230000003190 augmentative effect Effects 0.000 description 1
- 239000012472 biological sample Substances 0.000 description 1
- MCMSPRNYOJJPIZ-UHFFFAOYSA-N cadmium;mercury;tellurium Chemical compound [Cd]=[Te]=[Hg] MCMSPRNYOJJPIZ-UHFFFAOYSA-N 0.000 description 1
- 229910052791 calcium Inorganic materials 0.000 description 1
- 244000213578 camo Species 0.000 description 1
- 235000009120 camo Nutrition 0.000 description 1
- 235000019994 cava Nutrition 0.000 description 1
- 239000004568 cement Substances 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 229930002875 chlorophyll Natural products 0.000 description 1
- 235000019804 chlorophyll Nutrition 0.000 description 1
- ATNHDLDRLWWWCB-AENOIHSZSA-M chlorophyll a Chemical compound C1([C@@H](C(=O)OC)C(=O)C2=C3C)=C2N2C3=CC(C(CC)=C3C)=[N+]4C3=CC3=C(C=C)C(C)=C5N3[Mg-2]42[N+]2=C1[C@@H](CCC(=O)OC\C=C(/C)CCC[C@H](C)CCC[C@H](C)CCCC(C)C)[C@H](C)C2=C5 ATNHDLDRLWWWCB-AENOIHSZSA-M 0.000 description 1
- 238000007635 classification algorithm Methods 0.000 description 1
- 239000012141 concentrate Substances 0.000 description 1
- 230000001276 controlling effect Effects 0.000 description 1
- 238000001816 cooling Methods 0.000 description 1
- 235000005822 corn Nutrition 0.000 description 1
- 235000012605 creosote bush Nutrition 0.000 description 1
- 238000005520 cutting process Methods 0.000 description 1
- 230000002559 cytogenic effect Effects 0.000 description 1
- 208000002925 dental caries Diseases 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000001066 destructive effect Effects 0.000 description 1
- GUJOJGAPFQRJSV-UHFFFAOYSA-N dialuminum;dioxosilane;oxygen(2-);hydrate Chemical compound O.[O-2].[O-2].[O-2].[Al+3].[Al+3].O=[Si]=O.O=[Si]=O.O=[Si]=O.O=[Si]=O GUJOJGAPFQRJSV-UHFFFAOYSA-N 0.000 description 1
- 238000009792 diffusion process Methods 0.000 description 1
- 238000007598 dipping method Methods 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 239000000428 dust Substances 0.000 description 1
- 230000005274 electronic transitions Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000001917 fluorescence detection Methods 0.000 description 1
- 238000001215 fluorescent labelling Methods 0.000 description 1
- 230000005251 gamma ray Effects 0.000 description 1
- 239000007789 gas Substances 0.000 description 1
- 230000002068 genetic effect Effects 0.000 description 1
- 239000010442 halite Substances 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 238000010438 heat treatment Methods 0.000 description 1
- 230000036571 hydration Effects 0.000 description 1
- 238000006703 hydration reaction Methods 0.000 description 1
- 238000010191 image analysis Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000001727 in vivo Methods 0.000 description 1
- 208000015181 infectious disease Diseases 0.000 description 1
- 238000003331 infrared imaging Methods 0.000 description 1
- 150000002484 inorganic compounds Chemical class 0.000 description 1
- 229910010272 inorganic material Inorganic materials 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- JEIPFZHSYJVQDO-UHFFFAOYSA-N iron(III) oxide Inorganic materials O=[Fe]O[Fe]=O JEIPFZHSYJVQDO-UHFFFAOYSA-N 0.000 description 1
- NLYAJNPCOHFWQQ-UHFFFAOYSA-N kaolin Chemical compound O.O.O=[Al]O[Si](=O)O[Si](=O)O[Al]=O NLYAJNPCOHFWQQ-UHFFFAOYSA-N 0.000 description 1
- 229910052622 kaolinite Inorganic materials 0.000 description 1
- 238000009533 lab test Methods 0.000 description 1
- 210000000265 leukocyte Anatomy 0.000 description 1
- 230000031864 metaphase Effects 0.000 description 1
- 230000000394 mitotic effect Effects 0.000 description 1
- 229910052901 montmorillonite Inorganic materials 0.000 description 1
- 210000000214 mouth Anatomy 0.000 description 1
- 239000003921 oil Substances 0.000 description 1
- 238000007557 optical granulometry Methods 0.000 description 1
- 238000000399 optical microscopy Methods 0.000 description 1
- 150000002894 organic compounds Chemical class 0.000 description 1
- 230000010355 oscillation Effects 0.000 description 1
- 230000001151 other effect Effects 0.000 description 1
- 238000010422 painting Methods 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 230000029553 photosynthesis Effects 0.000 description 1
- 238000010672 photosynthesis Methods 0.000 description 1
- 230000000704 physical effect Effects 0.000 description 1
- 239000000049 pigment Substances 0.000 description 1
- 230000010287 polarization Effects 0.000 description 1
- 238000012827 research and development Methods 0.000 description 1
- 230000003938 response to stress Effects 0.000 description 1
- 229940043267 rhodamine b Drugs 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 230000009758 senescence Effects 0.000 description 1
- 230000035807 sensation Effects 0.000 description 1
- 229910052710 silicon Inorganic materials 0.000 description 1
- 239000010703 silicon Substances 0.000 description 1
- 235000002639 sodium chloride Nutrition 0.000 description 1
- 239000007921 spray Substances 0.000 description 1
- 206010041823 squamous cell carcinoma Diseases 0.000 description 1
- MPLHNVLQVRSVEE-UHFFFAOYSA-N texas red Chemical compound [O-]S(=O)(=O)C1=CC(S(Cl)(=O)=O)=CC=C1C(C1=CC=2CCCN3CCCC(C=23)=C1O1)=C2C1=C(CCC1)C3=[N+]1CCCC3=C2 MPLHNVLQVRSVEE-UHFFFAOYSA-N 0.000 description 1
- 238000002834 transmittance Methods 0.000 description 1
- WFKWXMTUELFFGS-UHFFFAOYSA-N tungsten Chemical compound [W] WFKWXMTUELFFGS-UHFFFAOYSA-N 0.000 description 1
- 229910052721 tungsten Inorganic materials 0.000 description 1
- 239000010937 tungsten Substances 0.000 description 1
- 229910052720 vanadium Inorganic materials 0.000 description 1
- 238000001429 visible spectrum Methods 0.000 description 1
- 239000002351 wastewater Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01J—MEASUREMENT OF INTENSITY, VELOCITY, SPECTRAL CONTENT, POLARISATION, PHASE OR PULSE CHARACTERISTICS OF INFRARED, VISIBLE OR ULTRAVIOLET LIGHT; COLORIMETRY; RADIATION PYROMETRY
- G01J3/00—Spectrometry; Spectrophotometry; Monochromators; Measuring colours
- G01J3/12—Generating the spectrum; Monochromators
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01J—MEASUREMENT OF INTENSITY, VELOCITY, SPECTRAL CONTENT, POLARISATION, PHASE OR PULSE CHARACTERISTICS OF INFRARED, VISIBLE OR ULTRAVIOLET LIGHT; COLORIMETRY; RADIATION PYROMETRY
- G01J3/00—Spectrometry; Spectrophotometry; Monochromators; Measuring colours
- G01J3/12—Generating the spectrum; Monochromators
- G01J3/1256—Generating the spectrum; Monochromators using acousto-optic tunable filter
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01J—MEASUREMENT OF INTENSITY, VELOCITY, SPECTRAL CONTENT, POLARISATION, PHASE OR PULSE CHARACTERISTICS OF INFRARED, VISIBLE OR ULTRAVIOLET LIGHT; COLORIMETRY; RADIATION PYROMETRY
- G01J3/00—Spectrometry; Spectrophotometry; Monochromators; Measuring colours
- G01J3/12—Generating the spectrum; Monochromators
- G01J3/26—Generating the spectrum; Monochromators using multiple reflection, e.g. Fabry-Perot interferometer, variable interference filters
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01J—MEASUREMENT OF INTENSITY, VELOCITY, SPECTRAL CONTENT, POLARISATION, PHASE OR PULSE CHARACTERISTICS OF INFRARED, VISIBLE OR ULTRAVIOLET LIGHT; COLORIMETRY; RADIATION PYROMETRY
- G01J3/00—Spectrometry; Spectrophotometry; Monochromators; Measuring colours
- G01J3/28—Investigating the spectrum
- G01J3/2823—Imaging spectrometer
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01J—MEASUREMENT OF INTENSITY, VELOCITY, SPECTRAL CONTENT, POLARISATION, PHASE OR PULSE CHARACTERISTICS OF INFRARED, VISIBLE OR ULTRAVIOLET LIGHT; COLORIMETRY; RADIATION PYROMETRY
- G01J3/00—Spectrometry; Spectrophotometry; Monochromators; Measuring colours
- G01J3/28—Investigating the spectrum
- G01J3/44—Raman spectrometry; Scattering spectrometry ; Fluorescence spectrometry
- G01J3/4406—Fluorescence spectrometry
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01J—MEASUREMENT OF INTENSITY, VELOCITY, SPECTRAL CONTENT, POLARISATION, PHASE OR PULSE CHARACTERISTICS OF INFRARED, VISIBLE OR ULTRAVIOLET LIGHT; COLORIMETRY; RADIATION PYROMETRY
- G01J3/00—Spectrometry; Spectrophotometry; Monochromators; Measuring colours
- G01J3/28—Investigating the spectrum
- G01J3/45—Interferometric spectrometry
- G01J3/453—Interferometric spectrometry by correlation of the amplitudes
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/62—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light
- G01N21/63—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light optically excited
- G01N21/64—Fluorescence; Phosphorescence
- G01N21/6428—Measuring fluorescence of fluorescent products of reactions or of fluorochrome labelled reactive substances, e.g. measuring quenching effects, using measuring "optrodes"
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/62—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light
- G01N21/63—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light optically excited
- G01N21/64—Fluorescence; Phosphorescence
- G01N21/645—Specially adapted constructive features of fluorimeters
- G01N21/6456—Spatial resolved fluorescence measurements; Imaging
- G01N21/6458—Fluorescence microscopy
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
- G01N33/48—Biological material, e.g. blood, urine; Haemocytometers
- G01N33/50—Chemical analysis of biological material, e.g. blood, urine; Testing involving biospecific ligand binding methods; Immunological testing
- G01N33/5005—Chemical analysis of biological material, e.g. blood, urine; Testing involving biospecific ligand binding methods; Immunological testing involving human or animal cells
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
- G01N33/48—Biological material, e.g. blood, urine; Haemocytometers
- G01N33/50—Chemical analysis of biological material, e.g. blood, urine; Testing involving biospecific ligand binding methods; Immunological testing
- G01N33/53—Immunoassay; Biospecific binding assay; Materials therefor
- G01N33/569—Immunoassay; Biospecific binding assay; Materials therefor for microorganisms, e.g. protozoa, bacteria, viruses
- G01N33/56966—Animal cells
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
- G01N33/48—Biological material, e.g. blood, urine; Haemocytometers
- G01N33/50—Chemical analysis of biological material, e.g. blood, urine; Testing involving biospecific ligand binding methods; Immunological testing
- G01N33/58—Chemical analysis of biological material, e.g. blood, urine; Testing involving biospecific ligand binding methods; Immunological testing involving labelled substances
- G01N33/582—Chemical analysis of biological material, e.g. blood, urine; Testing involving biospecific ligand binding methods; Immunological testing involving labelled substances with fluorescent label
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/88—Image or video recognition using optical means, e.g. reference filters, holographic masks, frequency domain filters or spatial domain filters
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/13—Satellite images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/64—Three-dimensional objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/69—Microscopic objects, e.g. biological cells or cellular parts
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01J—MEASUREMENT OF INTENSITY, VELOCITY, SPECTRAL CONTENT, POLARISATION, PHASE OR PULSE CHARACTERISTICS OF INFRARED, VISIBLE OR ULTRAVIOLET LIGHT; COLORIMETRY; RADIATION PYROMETRY
- G01J3/00—Spectrometry; Spectrophotometry; Monochromators; Measuring colours
- G01J3/28—Investigating the spectrum
- G01J2003/2866—Markers; Calibrating of scan
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01J—MEASUREMENT OF INTENSITY, VELOCITY, SPECTRAL CONTENT, POLARISATION, PHASE OR PULSE CHARACTERISTICS OF INFRARED, VISIBLE OR ULTRAVIOLET LIGHT; COLORIMETRY; RADIATION PYROMETRY
- G01J3/00—Spectrometry; Spectrophotometry; Monochromators; Measuring colours
- G01J3/02—Details
Definitions
- the present invention relates to remote sensing. More particularly, the present invention relates to a method for classification of remote scenes by decorrelation statistical analysis of spectra systematically derived from the scenes. The invention further relates to hardware for such classification, the hardware is dedicated or tunable according to parameters derived from the decorrelation statistical analysis.
- an image provides two sources of information.
- One is shape, and the other is intensity of tones and colors.
- the analysis of shape is a manual procedure and is still the state-of-the-art for getting terrain information about physical properties and conditions in relation to engineering applications such as site selection and evaluation, and to military applications such as cross-country movement.
- the second source of information is the analysis and evaluation of the colors, tones, and their intensities associated with a given pattern element.
- the need to get information by observation from afar is old.
- the techniques employed are very diverse, ranging from watching events from a hill, to probing the depths of space with the latest in technology.
- the workhorse for obtaining terrain information is electromagnetic radiation, be it reflected, emitted, or luminesced. Such radiation comes from natural sources, such as the sun, as well as from man-made sources. How it is measured is of equal diversity and includes sensors such as the eye, photographic emulsions, photo cells, antennae, charge coupled devices, thermistors, etc. Most often, the results from remote collection systems are presented as an image, or an assembly of images, wherein each image portrays the terrain in a different part of the electromagnetic spectrum.
- Targeting refers to the detection and, hopefully, the identification of specific features, items, or conditions.
- the target must have some characteristic that differs from its background and which cannot be confused with any other feature in the field of view. For point targets, this is seldom the case.
- the prediction of detectability is based on physics, and more often than not, this can be dealt with. But, more often than not, identification is iffy, mostly because of signals that, in the spectral band of use, look like the target.
- passive systems such as thermal infrared and thermal microwave, these issues are more complex than for the active systems. In general, the more cluttered the scene, for any type of sensor, the more difficult the identification of point source targets.
- Detection based on shape, size, and arrangement includes examples such as roads, airports, dams, vehicles, crop/field patterns, structures, and urban areas; and such are usually easier to identify by manual analysis. In addition to shape, differences can also be based in color, spectral reflectance, spectral emittance (temperature), luminescence, acoustical reflectance and emittance, magnetic fields, etc.
- Application examples include road type (asphalt or cement concrete), diseased vegetation, stressed vegetation, flood boundaries, wetland areas, thermal springs, thermal plumes for power plants, oil slicks, hot spots in burned areas, alteration zones, number of lakes in a region, camouflaged sites, and military units and equipment.
- a camouflaged target may resemble the terrain visually, photographically, and thermally, but show a mismatch in the radar bands; or, match in the radio frequencies, but not in the thermal infrared; or match in radar and thermal infrared, but show a mismatch in its solar spectral signature.
- EM electromagnetic
- locations within it can be specified by numbers for frequency (cycles per second) and wavelength (Angstroms, nanometers, micra, feet, miles, etc.), or by words.
- Words that designate frequency locations are radio terms such as Low Frequency (LF), High Frequency (HF), Very High Frequency (VHF), Ultra High Frequency (UHF), and Extraordinary High Frequency (EHF), at which point they ran out of superlatives.
- LF Low Frequency
- HF High Frequency
- VHF Very High Frequency
- UHF Ultra High Frequency
- EHF Extraordinary High Frequency
- amateur radio operators use the terms, 1 meter band, 10 meter band, etc.
- the visible part of the EM spectrum covers the approximate wavelength range of 400-700 nanometers (nm). It is not a sharp cutoff at the long wavelength end, and many people can perceive photons with wavelengths well beyond 700 nm.
- the 400-500 nm band is called blue, the 500-600 nm band green, and the 600-700 nm band red. Wavelengths shorter than 400 nm fall in the ultraviolet region, and wavelengths longer than 700 nm are in the infrared domain, which extends out to a wavelength of 1 mm.
- the microwave region When electromagnetic energy falls on a material, be it gas, liquid, or solid, several things can happen. What happens depends on the electrical properties of the materials, i.e., the index of refraction, or dielectric constant. In turn, this is a function of the oscillation frequency of the incoming electric field. More exactly, what happens depends on changes in the electrical properties between the medium the radiation it is in, and the medium it is entering. If, in going from one medium to another, the radiation encounters a change in electrical properties that takes place in a distance less than its characteristic wavelength, then something must happen to that radiation - it cannot continue as it was. It will undergo, singly or in combination, reflection (scattering), refraction, or absorption.
- the recording of the reflected component of radiation is the most common form of remote sensing, and can involve the sun, lasers, and radio frequencies.
- the sun is the most frequently used source, and in the reflected solar spectrum (0.4-2.5 micra), supports sensors such as cameras, the Landsat Multispectral Scanner (MSS) and Thematic Mapper (TM) bands 1, 2, 3, 4, 5, and 7 (band 6 being a thermal infrared band), SPOT, and the hyperspectral systems such as the Airborne Visible Infrared Imaging Spectrometer (AVIRIS).
- the tone differences that define boundaries of shapes, soil changes, and the highlight and shadow tones are due to different amounts of radiation reflected to the sensor from the various surfaces.
- the amount of radiation reflected from a surface depends on the wavelength band being used, its angle of incidence with the surface, the orientation of the sensor in relation to the surface and the illuminant, the material's molecular composition, and the surface structure.
- Molecular composition sets the stage, and as we progress to remote sensors that record reflected energy in narrower and narrower bands, we become more interested in the details of the interaction of electromagnetic energy with matter.
- a surface that reflects uniformly across the visible and photographic parts of the EM spectrum is called a neutral surface - i.e., it does not treat any one wavelength different from any other.
- no neutral material is a perfect reflector, i.e., 100% reflectance throughout the spectrum.
- it a perfect absorber i.e., 0% reflectance. If it has a high reflectance, it is called white. If it has a low reflectance, it is called black. In between it is called gray.
- a blue surface reflects blue light, and absorbs green and red
- a yellow surface absorbs blue
- a green surface reflects green light, but absorbs blue and red.
- three surfaces, e.g., blue, green, and red reflect the same amount of photons - i.e., the intensity from the blue surface is the same as the green, which is the same as the red - then a black and white photograph would show them as the same gray tone - they would be indistinguishable.
- a normal color film is a multiband system. It has three bands, or channels, that separately document the intensities in the blue, green, and red bands, and displays the results as a color composite - showing that there are three different objects in the scene.
- healthy vegetation absorbs blue light, reflects some of the photons in the green band, and absorbs photons in the red band - thus, we see healthy leafy vegetal material as green, and it would be recorded as such in a normal color film. Healthy leafy vegetation is much more reflective in the infrared part of the spectrum, than in the visible; which is why, in black and white infrared photography, most vegetation shows as very bright tones. Vegetation stress may be detected at the 1.4 and 1.9 micra regions which reflect water absorption. As a plant loses water, these bands become shallower, and thereby provide another indicator of vegetation stress.
- the infrared brightness of healthy vegetation is a useful characteristic for military targeting.
- WW II Eastman Kodak was asked to develop a film that could separate green painted camouflaged targets from the green surrounds of vegetal material. It turns out that, though a visual match, green paint is much less reflective in the infrared, than is vegetation.
- color films such as Kodachrome, were sensitive primarily to photons in the visible part of the spectrum.
- Eastman Kodak developed the needed film, which was known as Camouflage Detection film, or CD film.
- this film would portray green paint as green, and healthy vegetation as red.
- the emulsion that had the extended sensitivity was the one sensitive to bluelight, and which was coupled to the cyan dye. Increasing its sensitivity to infrared, out to 900 nm, greatly increased its sensitivity to the more energetic photons at double the frequency, or half the wavelength, i.e., 400-450 nm.
- the emulsion was about ten times more sensitive to blue light than to infrared.
- a yellow filter was required to prevent blue light from entering the emulsion.
- such a filter passes two thirds of the visible spectra, i.e., green and red, which causes the visual sensation we call yellow. Because the filter rejects the blue third, it is also called a minus-blue filter. Examples include Wratten 11, 12, and 13. With the blue light eliminated, the red response could be attributed to the vegetation - and any green painted object stuck out in stark contrast as green.
- the colors in an infrared photograph are not colors that eyes see in the scene, such a photograph is also called a false color image.
- One response of stressed vegetation is loss of the bright infrared reflectance.
- the stress can be induced by drought, flooding, chemical sprays, senescence, by biological infections such as the rust and wilt, or by infestations such as gypsy moths.
- the loss of infrared reflectance provides the basis for detecting the presence of the stress, mapping its extent, and monitoring an area for change.
- camouflage was, and still is, frequently made by cutting branches of trees to lay over the site - and this is effective, but only for a relatively short time.
- One response of the cut branch is that the leafy material loses its infrared reflectance faster than it loses its green reflectance. Even though looking equally green to the eye, the CD film shows the reflectance loss of the damaged vegetation as dark tones of red/green and green.
- EIR Ektachrome Infrared film
- Normal color film has three spectral bands, or recording channels, i.e., blue, green, and red, which are analogous to bands 1, 2 and 3 of the Landsat Thematic Mapper.
- the color combines the three channels into a normal color image - a color composite.
- Adding infrared sensitivity makes four channels - blue, green, red, and infrared-analogous to Landsat TM bands 1, 2, 3 and 4.
- Landsat TM With Landsat TM, the selection is made electronically. Of the seven bands available, one can select band 2, 3 and 4, couple them to blue, green, and red guns and create a false color composite image that is similar to the image from the EIR film with a yellow filter.
- the reflectance of earth materials increases as one goes to longer wavelengths.
- most surfaces have a similar low reflectance. This is why photographs taken in ultraviolet, and in blue light tend to be flat - there is little contrast. Reflectance steadily increases going from blue, to green, to red, and to infrared. It is in the infrared region that brightness, contrast, and other interesting spectral features begin to develop.
- the reflectance characteristics of playa surface materials such as calcite, halite, montmorillonite, kaolinite and gypsum have common absorption bands. Strong water absorption bands at 1400 nm and 1900 nm are apparent for some of the materials.
- these broad bands do not represent a single absorption, but a collection of narrower bands.
- free water has several absorption bands, including ones at 1350 nm, 1380 nm, and 1460 nm. Unless the spectrometer has a narrow spectral bandpass, these closely adjacent bands cannot be resolved - they blend into one larger band.
- the 1400 nm band for gypsum shows structure within it that suggest the presence of other bands.
- a record that shows both bands, i.e., 1400 nm and 1900 nm indicates undissociated water such as water of hydration, or water trapped in the lattice.
- Molecular water which is an important component in gypsum, has other absorption bands known as overtones and combinatorial tones.
- the reflected component can be greatly affected by other factors, an important one being surface structure.
- Surface structure can be defined in terms of wavelength of the radiation concerned. For a surface to be considered a high quality reflector, or mirror, it must be flat to within about a quarter of a wavelength of the radiation to be used. Under this condition, the bulk of the incoming radiation is reflected at an angle that equals the angle of incidence. This is called specular reflection. If the topographic variations of the surface exceed this, then proportionally more of the energy is scattered in other directions, i.e., the diffuse component gets smaller.
- the radiation wavelengths fall between 400 nm and 2500 nm, or between 0.4 and 2.5 micra.
- most surfaces have relief variations in excess of this, and are diffuse scatters, e.g., leaves, bark, soil, rocks, concrete, etc. Consequently, when viewing terrain from some point in space, the image tones can vary, even be reversed, as a function of viewing angle in relation to the sun.
- the sensor receives both the diffuse and specular reflection components. Down-sun, the sensor receives diffuse minus specular, i.e., only the backscatter. If the surface variations have an orderly structure, the image patterns are influenced by sun azimuth as well.
- the ridges and furrows of a freshly plowed field have relief on the order of 15 to 20 cm. Furthermore, it is an ordered relief, i.e., a pattern of parallel lines. If the sun's rays are parallel to the ridge pattern, the surface receive about the same amount of illumination and the image of the field would have an overall uniform tone. If the sun's rays are perpendicular to the ridge pattern, the field shows as a parallel series of highlights and shadows. X- and C-band radars would show a similar display. For P-band radar (100 cm wavelength), however, the plowed field would be a speculum, or mirror, and would be a dark area, an area of no return, or radar loss, from any angle of illumination.
- Radiation absorbed by a material leads to other effects. For one, absorbed photons increase the internal energy, or temperature, of the material, which, in turn, increases the quantity and alters the wavelength distribution of the thermally emitted radiation, both infrared and microwave. This is the most common outcome of absorption, and is the basis for thermal, or passive, remote sensing in either the infrared or the microwave domain.
- absorbed sunlight heats the terrain. During the night the terrain cools by radiating into space, and would continue to cool except that the sun comes up and renews the heating cycle.
- Thermal infrared techniques are associated with two wavelength regions for which the atmosphere is transparent, i.e., atmospheric windows. These are the 3.5- 5.5 and 8-14 micra bands. Although thermal techniques are considered by many to be outside the hyperspectral domain, we include them for practical reasons. First, characteristics of thermal images and the events that influence them, are so different from images formed with reflected radiation that they deserve special mention. Second, terrain analysts in DoD, must be able to provide information from any imagery source. Third, thermal systems provide a different type of image, which is important for targeting. Fourth, these systems are available, and being used. Examples include the Landsat TM band 6 (10.4-12.5 micra wavelength), and the Thermal Infrared Mapping system (TIMS), which has six bands in the 8-14 micra range.
- GPS Thermal Infrared Mapping system
- This signal is amplified, displayed on a cathode ray tube and recorded on magnetic tape, or on photographic emulsions via a modulated glow tube, or other device.
- the images are printed so that light tones represent warmer surfaces, and dark tones represent cooler surfaces.
- the amount of energy emitted from a surface, and its wavelength distribution, depend on temperature.
- the amount is equal to the fourth power of the absolute temperature multiplied by the emissivity.
- the wavelength distribution curve will have a maximum intensity as a specific wavelength known as lambda sub-max ( ⁇ max).
- ⁇ max shifts to shorter wavelength, e.g., at -150 °C it is about 23.2 micra, at 0 °C it is about 10.5 micra, and at 100 °C it is about 7.7 micra.
- ⁇ max is about 0.5 micra.
- the short wavelength edge of the distribution curve includes higher energy levels, i.e., it goes to shorter wavelengths, and the area under the curve (total energy emitted) gets rapidly larger.
- emissivity constant, a warmer body will emit more energy at all wavelengths than will a cooler body, and will incorporate a short wavelength increment denied the cooler material.
- the wavelength of peak emission is in the 8-14 micra band, which is one of the atmospheric windows. About 40 % of the energy is emitted in this band, and about 3 % in the 3.0 - 5.5 micra band.
- the 8-14 micra band is the preferred choice.
- the wavelength of peak emission enters the 3.0 - 5.5 micra band. For detecting hot targets, this band is the preferred choice.
- the signal/noise ratio, or the target/background contrast is much greater here.
- Emissivity which is wavelength dependent, denotes how good an absorber, or emitter, a material is. Molecules emit energy only at those wavelengths they can absorb. A perfect absorber, or emitter, has an emissivity of 1. A material that absorbs 50% of the incoming radiation, and reflects 50%, has an emissivity of 0.5.
- This layer is an interface between the material below, and the atmosphere above, and is readily influenced by events on both sides.
- energy transfer is associated with conduction and, in some cases, diffusion.
- atmospheric variables take over. These can quickly, and radically, alter the radiation characteristics of the surfaces - eliminating, subduing, or increasing thermal contrast.
- thermal contrast is influenced by a variety of diurnal and seasonal variations in climatic and meteorological factors, such as wind, atmospheric pressure, dew, rain, humidity, incoming space radiation, etc. Wind can override subsurface conductive events and imprint its own temperature regime, which can create confusing thermal patterns in the form of wind shadows - i.e., surfaces in the lee can be much warmer, or cooler, than surfaces exposed to the wind.
- Objects sticking up into the air take on the temperature characteristics of the air. At nighttime, when the air is warmer than the ground, which is cooling by radiation loss, these objects will appear as hotspots. As the night progresses, the air layer, cooled by contact with the cooler ground, becomes thicker, and sequentially cools taller objects that project up into it. By late night the hotspots disappear, except for the tallest trees, or for natural or artificially maintained heat sources.
- cooler air is more dense, and being more dense, it flows downslope to settle in the lows, which show as darker tones in the thermal imagery. Darker tones in the lows can also be caused by moist soil, and can lose heat faster. As a result, cool air drainage into the lows is sometimes mistaken for moist soils.
- atmospheric pressure changes are critical. First, the interior temperatures are usually fairly constant, and are warmer or cooler than the ground surfaces at some time of the day. Also, infrared cannot penetrate the overburden to reveal the presence of the void.
- the best time to fly is when the void is exhaling, and the outpouring flow of warm air through various openings brings the temperature of the surrounds to above ambient, and they become detectable - i.e., one detects the openings, not the void. Such an outpouring can occur only when the atmospheric pressure is less than the air pressure in the void. Thus, the time to fly is on a descending pressure front.
- a second effect of photon absorption is luminescence.
- An example is the emission of visible light from minerals when they are illuminated with "black" light, or ultraviolet radiation.
- Luminescence is an emission of radiation due to electronic transitions, and there are two kinds - fluorescence, which occurs from an excited single state, and phosphorescence, which results from an excited triple state. A distinction can also be made on the basis of time - i.e., how long does the light last after the excitation energy is turned off? This is called the decay time. In fluorescence, the decay time is very short, ranging from 10 ⁇ 9 to 10 seconds. For example, the fluorescence decay time of rhodamine B in water is about 2.5 nanoseconds (ns). In phosphorescence, the decay time is longer - sometimes much longer. Calcium sulf ⁇ de, for example, can continue to glow for several hours after the excitation illumination is turned off.
- Luminescent techniques require an energy source to excite, or raise the electrons to higher energy levels, and darkness in order to detect the luminesced photons.
- the sun meets both of these needs - as does the nighttime use of lasers.
- the sun does not emit a continuous spectrum, i.e., energy at all wavelengths, or frequencies. Although such is generated in the hot core, electronic absorption by elements and ionized atoms in the cooler envelope greatly reduces the intensities of many of the frequencies.
- the sun is examined with a good spectroscope, one finds that there are gaps - many gaps - wherein energy is greatly reduced, or absent. These gaps of darkness are narrow in bandwidth, so narrow they are called lines - specifically, Fraunhofer Lines in honor of their discoverer.
- the spectral bandwidth of these lines are measured in Angstrom units.
- the ultraviolet, visible, and near infrared portion of the solar spectrum contain over 30,000 Fraunhofer Lines, or lines of darkness. These lines provide the darkness needed for detection of luminescence, and the sun's radiation provides the excitation energy on the short wavelength side of the lines.
- a sensor system that can look at both the sun and the earth's surface with detectors sensitive to energy in these dark lines can detect the presence of luminescence photons from the earth's surface, or target. If, from the target, it detects a certain intensity in a dark line band, it cannot be reflected solar energy because such is not coming from the sun. The fill-in must, therefore, be due to luminesced photons from the target. Such is the function of the Fraunhofer Line Discriminator (FLD). Such things as temperature and pH can alter the characteristics of the luminesced photons. Some materials that are under constant, or steady-state illumination, give a different luminescence signal after an hour or two, than they do to an instant measurement immediately after excitation.
- FLD Fraunhofer Line Discriminator
- luminescence Another characteristic of luminescence is that for any specific wavelength of excitation, there is an emission spectra that can take place over a fairly broad wavelength band, and the decay times of the longer wavelengths can be considerably longer than those of the shorter wavelengths. The total is still a very short time. In laboratory experiments, decay time spectra have shown links to material types and conditions. Whether or not such has application in remote sensing remains to be determined. Because all materials reflect, absorb, or emit photons in ways characteristic of their molecular makeup, a high resolution trace of the intensity of the transmitted, reflected, emitted, or luminesced radiation versus wavelength forms a graphical record unique to a given material. Different materials cannot have identical spectral wave shapes of reflectance, emittance, and luminescence.
- Photographic emulsions were the earliest of the sensors to document landscape scenes, and human activities. By the late 1800's emulsions went airborne, via balloons and kites, to replace the observer and his notepad for recording terrain characteristics and military items of interest. By WW I, cameras were in airplanes and routinely involved in reconnaissance and targeting. It was recognized early on that if one could sample different wavelength of radiation, and compare them, one would have a better chance of detecting targets, as well as noting changes in the landscape - a multiband concept, although not called that, was now in place.
- the first steps were taken in the late 1940's and the 1950's when the Army, along other groups, divided the photographic portion of the electromagnetic spectrum, 400 to 900 nm, into narrower bandpasses by means of various combinations of photographic emulsions and filters.
- the goal was to improve techniques for detecting targets and mapping conditions such as camouflage, vegetation type, vegetation stress, soil moisture, flood damage, wetland boundaries, to name a few, and the term multiband photography came into being to describe these efforts.
- Camouflage Detection film, and its improvement into Ektachrome Infrared film is one example of a successful film filter combination, or multiband approach which later passed into the digital domain of Landsat as the False Color Composite.
- the bandpasses were still broad, however, ranging from 60 to 100 nm. Nevertheless, multiband photography had applications to some forms of targeting, and change detection.
- AIS Airborne Imaging Spectrometer
- JPL Jet propulsion Laboratory
- the AIS records reflected solar energy in some 128 channels, or images, within the 1.2 - 2.4 micra region of the spectrum and with a spectral bandwidth for each channel of less than 10 nm.
- the AIS evolved into the AVIRIS with some 220 raw data channels, or images, within the 0.4 - 2.45 micra portion of the spectrum. Resampling gives 210 spectral bands of radiometrically calibrated data.
- the instantaneous field of view (IFOV) is 1 milliradian, or about 10 meters at operational altitude.
- Each image is a record of the intensity of reflected sunlight within a spectral bandwidth of less than 10 nm.
- the intensity values of the 210 channels, for any given picture element (pixel) can be called up and sequentially displayed along the wavelength axis, as a spectrophotometric trace, i.e., radiometric intensity versus wavelength.
- these systems are called hyperspectral, to differentiate them from the broad band systems, e.g., MSS, TM, SPOT, etc.
- Systems are also being developed that can operate in even narrower bands, i.e., the sub-nanometer range, for working with gaseous emissions and absorptions. These are called ultraspectral systems.
- hyperspectral refers to a multiplicity of recording channels that have relatively narrow bandwidths.
- AIS Shuttle Imaging Spectrophotometer Experiment
- HIRIS High Resolution Imaging Spectrometer
- the atmosphere absorbs many wavelength components of the incoming sunlight, as well as of the reflected energy en route to the sensor, corrections are needed for many targets. If one is interested in vegetation, this involved the depth and shapes of water absorption bands. Because water vapor is a component of the atmosphere, the analyst does not know how much of the depth and shape of those water bands is due to atmospheric absorption, and how much is due to vegetation absorption. If corrections can be made to remove the atmospheric component via available models such as LowTran, then the residuum can be attributed to plant water.
- the question is easier to answer.
- the geologist can get by with perhaps 30 to 40 bands.
- the agriculturist can get by with perhaps 20 bands, only a portion of which overlap the geologists' needs.
- the army with its interest in terrain, targeting, and intelligence, has need for information about identities, and properties associated with vegetation, soils, rocks, minerals, and cultural objects including camouflage. Perhaps reduction can be made - perhaps there are bands that have no use for anybody - but, it is too early for declarations.
- an imaging spectrometer provides two domains of information for evaluation-image patterns and spectral patterns. From the standpoint of terrain information in terms of materials identities and conditions, potential for dust generation, location of engineering materials, engineering site selection and evaluation, probable location of ground water, surface waste disposal, etc., the manual analysis of stereo imagery is still state-of- the-art. For example, an area can be covered with a vegetative mantle of grass and trees, and all that the spectral data will show will be reflectance traces of chlorophyll.
- the shapes of the landform and drainage can reveal that beneath the vegetal mantle rests a thinly interbedded series of limestones and shales dipping gently to the west, and with unstable colluvial materials on the lower slopes.
- imaging spectrometers provide only monoscopic imagery, so there is a reduction in the quantity and quality of information that can be derived on the basis of image pattern shapes; but these shapes are present, and they can make significant direct contributions to an analysis, as well as assist in the valuation of the spectral data.
- existing routines for combining bands to make color composite images such as Landsat, or the Coastal Zone color Scanner (CZCS), can be directly applied to hyperspectral data.
- the airborne imaging spectrometers are here, the spaceborne systems are in development, and the hyperspectral concept is sound.
- the issues to be resolved include: what are these systems suited for? - what are their advantages, disadvantages, and limitations? - and, how well will they work?
- Spectral data from imaging spectrometers can be evaluated on the basis of: shape of the overall curve, or portions of it; intensity differences at any selected wavelength range; wavelength location of absorption bands; and, depth and shape of absorption bands.
- To link these to identities and conditions requires an extensive computer library of field and laboratory measurements of spectral reflectance, luminescence, and emittance throughout the reflected solar, and thermal infrared portions of the spectrum - and the software to make the evaluations and comparisons.
- TEC has a spectral reflectance/luminescence data base of over 1,000 samples of soils, rocks, vegetation, and man-made materials.
- Such a library needs excellent documentation, because these measured values change with a variety of factors for any given surface, the molecular makeup determines the basic characteristics of absorption, reflectance luminescence, and emittance. These in turn, are modified by structure of the surface, and its orientation in relation to the sensor and to the illuminating source. For example, maintaining a constant field of view and a constant viewing angle, while measuring spectral reflectance at different sun angles and elevations, can result in variances of plus or minus 10 percent.
- vegetation can have smooth, crenelated, or wrinkled leaf surfaces, and the leaves and stems can have many different sizes and be arranged in many different ways. This means different highlight/shadow ratio, different amounts of transmitted and re-reflected infrared energy through the biomass, and different amounts of radiation reflecting up through the vegetation from the soil surface.
- the spectral signature of a fine textures soil can differ from that of a coarser textured soil.
- New leaves have a different spectral signature than older leaves, wet soil is different than the same soil when dry, a wheathered rock surface differs from a fresh surface.
- these are different chemical forms which gets back to the earlier statement that the molecular makeup of a surface establishes the basis of reflectance and absorptance.
- the spectral signature is further modified by climate, season, and meteorological variations. Changes in incoming short and long wave radiation from space, wind and atmospheric pressure greatly alter radiometric signatures, as well as target/background contrasts in thermal imagery.
- Multiplicity of measurements is necessary because there can be significant variation within any given class of targets, especially in field measurements. For example, one can measure 20 creosote bushes that look alike and are about the same size and age. But, the result will likely be 20 slightly different spectra- perhaps plus or minus 10% variance, or more, from a derived norm. The variations are mostly in intensity, not wavelength locations of absorption bands. Although the plants look alike, they are not identical - each has some variance in biomass, structure, openness, etc. these factors alter the characteristics of the energy reflected from the vegetal surfaces, as well as the characteristics of the contributing reflected soil component passing through, or reflected from the canopy. For current systems and typical target areas, the IFOV (10 meters for
- AVIRIS encompasses a mixture of surfaces, and the resulting spectral signature is a composite of individual signatures - which presents another problem in relation to digital analysis of spectra data.
- the present invention is directed at providing a system (hardware and software) which performs a measurement, with higher sensitivity and at higher speed, and encompassing a much smaller amount of data from the outset.
- the hardware does not require an interferometer, but only a number (N) of what is herein referred to as "decorrelation matched filters", which are placed in the path of the incoming light beam from the remote scenes to be measured.
- the filters may be of a fixed nature or tunable (AOTF or LCTF). In the latter case a single tunable filter is used to sequentially mimic the decorrelation matched filters under electronic control.
- a method for preparing a reference template for classification of remote scenes comprising the steps of (a) classifying a set of reference scenes via a conventional classification technique for obtaining a set of preclassified scenes; (b) using a spectral imager for measuring a spectral cube of the preclassified referenced scenes; (c) employing a decorrelation statistical method for extracting the spectral cube for decorrelated spectral data characterizing the reference scenes; and (d) using at least a part of the decorrelated spectral data for the preparation of the reference template for remote scenes classification.
- the principal component analysis includes the steps of (a) selecting k spectral slices for the spectral cube of the reference scenes; (b) calculating an average spectrum for each of the reference scenes; (c) stretching each of the average spectra for obtaining a stretched average spectrum for each of the reference scenes; (d) averaging the stretched average spectra for each of the reference scenes, for obtaining an ensemble average spectrum for each of the reference scenes; (e) calculating a k dimension eigen system for the ensemble average spectra and extracting N eigenvectors; (f) using the N eigenvectors for defining an N- dimension vector for each of the reference scenes; and (g) using the N-dimension vectors for preparing the reference template for the remote scenes classification.
- the principal component analysis further includes the step of (h) performing a spatial averaging procedure over all spectral slices.
- a method for remote scenes classification comprising the steps of (a) preparing a reference template for classification of the remote scenes via (i) classifying a set of reference scenes via a conventional classification technique for obtaining a set of preclassified reference scenes; (ii) using a first spectral imager for measuring a spectral cube of the preclassified reference scenes; (iii) employing a principal component analysis for extracting the spectral cube for decorrelated spectral data characterizing the reference scenes; and (vi) using at least a part of the decorrelated spectral data for the preparation of the reference template for remote scenes classification; (b) using a second spectral imager for measuring a spectral cube of analyzed remote scenes, such that a spectrum of each pixel in the remote scenes is obtained; (c) employing a decorrelation statistical method for extracting decorrelated spectral data characterizing the pixels; and (d) comparing at least a part of the decorrelated spectral data extracted from
- a method for remote scenes classification comprising the steps of (a) preparing a reference template for classification of remote scenes via (i) classifying a set of reference scenes via a conventional classification technique for obtaining a set of preclassified reference scenes; (ii) using a first spectral imager for measuring a spectral cube of the preclassified reference scenes; (iii) employing a decorrelation statistical method for extracting the spectral cube for decorrelated spectral data characterizing the reference scenes; and (iv) using at least a part of the decorrelated spectral data for the preparation of the reference template for the remote scenes classification; (b) using a second spectral imager for measuring a spectral cube of analyzed remote scenes, such that a spectrum of each pixel in the remote scenes is obtained; (c) projecting the spectrum of each of the pixels onto the decorrelated spectral data for obtaining a projected spectrum for each of the pixels; and (d) comparing the projected spectra
- a method for remote scenes classification comprising the steps of (a) preparing a reference template for classification of remote scenes via (i) classifying a set of reference scenes via a conventional classification technique for obtaining a set of preclassified reference scenes; (ii) using a first spectral imager for measuring a spectral cube of the preclassified reference scenes; (iii) employing a principal component analysis for extracting the spectral cube for decorrelated spectral data characterizing the reference scenes via (a) selecting k spectral slices for the spectral cube of the reference scenes; (b) calculating an average spectrum for each of the reference scenes; (c) stretching the average spectra for obtaining a stretched average spectrum for each of the reference scenes; (d) averaging the stretched average spectra for each of the reference scenes for obtaining an ensemble average spectrum for each of the reference scenes; (e) calculating a k dimension eigen system for the ensemble average spectra and extracting N eigen
- a method of calculating decorrelation matched filters for remote scenes classification the decorrelation matched filters being for extracting decorrelated spectral data from the remote scenes
- the method comprising the step of (a) obtaining decorrelated spectral data characterizing a set of reference scenes via (i) classifying the set of reference scenes via a conventional classification technique for obtaining a set of preclassified reference scenes; (ii) using a first spectral imager for measuring a spectral cube of the preclassified reference scenes; and (iii) employing a decorrelation statistical method for extracting the spectral cube for decorrelated spectral data characterizing the reference scenes; and (b) mathematically manipulating at least a part of the decorrelated spectral data for obtaining a mathematical description of the decorrelation matched filters.
- the decorrelated spectral data is obtained using a principal component analysis, which includes expressing each of the reference scenes by a linear combination of N
- a set of decorrelation matched filters for remote scenes classification the decorrelation matched filters being for extracting decorrelated spectral data from the remote scenes, the set comprising physical filters having shapes, the shapes following a mathematical description, the mathematical description being obtainable by (a) obtaining decorrelated spectral data characterizing a set of reference scenes via (i) classifying the set of reference scenes via a conventional classification technique for obtaining a set of preclassified reference scenes; (ii) using a first spectral imager for measuring a spectral cube of the preclassified reference scenes; and (iii) employing a decorrelation statistical method for extracting the spectral cube for decorrelated spectral data characterizing the reference scenes; and (b) mathematically manipulating at least a part of the decorrelated spectral data for obtaining the mathematical description of the decorrelation matched filters.
- a method of tuning a tunable filter for remote scenes classification the method renders the tunable filter to mimic a set of decorrelation matched filters, and is for extracting decorrelated spectral data from the remote scenes
- the method comprising the steps of (a) obtaining decorrelated spectral data characterizing a set of reference scenes via (i) classifying the set of reference scenes via a conventional classification technique for obtaining a set of preclassified reference scenes; (ii) using a first spectral imager for measuring a spectral cube of the preclassified reference scenes; and (iii) employing a decorrelation statistical method for extracting the spectral cube for decorrelated spectral data characterizing the reference scenes; (b) mathematically manipulating at least a part of the decorrelated spectral data for obtaining a mathematical description describing the set of decorrelation matched filters; and (c) sequentially tuning the tunable filter according to the mathematical description.
- the tunable filter is selected from the group consisting of AOTF and LCTF.
- a method for remote scenes classification comprising the steps of (a) preparing a reference template for classification of the remote scenes via (i) classifying a set of reference scenes via a conventional classification technique for obtaining a set of preclassified reference scenes; (ii) using a first spectral imager for measuring a spectral cube of the preclassified reference scenes; (iii) employing a decorrelation statistical method for extracting the spectral cube for decorrelated spectral data characterizing the reference scenes; and (iv) using at least a part of the decorrelated spectral data for the preparation of the reference template for the remote scenes classification; (b) calculating a mathematical description of decorrelation matched filters for classification of the remote scenes employing the reference template, the calculation being by mathematically manipulating at least a part of the decorrelated spectral data; (c) using the mathematical description of the decorrelation matched filters for manufacturing the decorrelation matched filters; (d) using the decorrelation matched filters
- a method for remote scenes classification comprising the steps of (a) providing a set of decorrelation matched filters for the remote scenes classification, the decorrelation matched filters being for extracting decorrelated spectral data from the remote scenes, the set including physical filters having shapes, the shapes following a mathematical description, the mathematical description being achieved by (i) preparing a reference template for classification of remote scenes via (a) classifying a set of reference scenes via a conventional classification technique for obtaining a set of preclassified reference scenes; (b) using a first spectral imager for measuring a spectral cube of the preclassified reference scenes; (c) employing a decorrelation statistical method for extracting the spectral cube for decorrelated spectral data characterizing the reference scenes; and (d) using at least a part of the decorrelated spectral data for the preparation of the reference template for the remote scenes classification; and (ii) mathematically manipulating at least a part of the decorrelated spectral data for obtaining the mathematical
- a method for remote scenes classification comprising the steps of (a) providing a tunable filter and tuning information for tuning the tunable filter so as to mimic a set of decorrelation matched filters, the tunable filter being for extracting decorrelated spectral data from the remote scenes, the tuning information being achieved by (i) preparing a reference template for classification of remote scenes via (a) classifying a set of reference scenes via a conventional classification technique for obtaining a set of preclassified reference scenes; (b) using a first spectral imager for measuring a spectral cube of the preclassified reference scenes; (c) employing a decorrelation statistical method for extracting the spectral cube for decorrelated spectral data characterizing the reference scenes; and (d) using at least a part of the decorrelated spectral data for the preparation of the reference
- a spectral decorrelation measurement apparatus for remote scenes classification by extracting decorrelated spectral data from the remote scenes
- the apparatus is connected to a telescope used to view the remote scenes
- the apparatus comprising (a) a detector; and (b) an optical system for transmitting electromagnetic radiation from the remote scenes onto the detector, the optical system including a set of decorrelating matched filters, the decorrelation matched filters being for extracting decorrelated spectral data from the remote scenes, the filters of the set of decorrelation matched filters having shapes, the shapes following a mathematical description, the mathematical description being calculated by (i) obtaining decorrelated spectral data characterizing a set of reference scenes via (a) classifying the set of reference scenes via a conventional classification technique for obtaining a set of preclassified reference scenes; (b) using a first spectral imager for measuring a spectral cube of the preclassified reference scenes; and (c) employing a decorrelation statistical method for extracting the spectral cube for decorrelated
- the decorrelation matched filters are arranged on a rotatable filter carrying element.
- a specfral decorrelation measurement apparatus for remote scenes classification by extracting decorrelated spectral data from the remote scenes
- the apparatus is connected to a telescope used to view the remote scenes
- the apparatus comprising (a) a detector; and (b) an optical system for transmitting electromagnetic radiation from the remote scenes onto the detector, the optical system including a tunable filter and a tuning device, the tuning device being for tuning the tunable filter, so that the tunable filter sequentially mimics a set of decorrelating matched filters, the decorrelation matched filters mimicked by the tunable filter being for extracting decorrelated spectral data from the remote scenes, the tuning of the tunable filter being calculated according to a mathematical description, the mathematical description being calculated by (i) obtaining decorrelated spectral data characterizing a set of reference scenes via (a) classifying the set of reference scenes via a conventional classification technique for obtaining a set of preclassified reference scenes; (b) using a first spectral imager for measuring a
- the spectral imager includes an element selected from the group consisting of a dispersion element, a filter, a tunable filter and an interferometer.
- the decorrelation statistical method is selected from the group consisting of principal component analysis, canonical variable analysis and singular value decomposition.
- the principal component analysis includes expressing each of the scenes as linear combinations of N eigenvectors.
- N is an integer greater than two.
- N is an integer greater than two and smaller than eight.
- the present invention successfully addresses the shortcomings of the presently known configurations by providing a system (hardware and software) which performs a measurement, with higher sensitivity and at higher speed, and encompassing a much smaller amount of data from the outset.
- the hardware does not require an interferometer, but only a number (N) of what is herein referred to as "decorrelation matched filters", which are placed in the path of the incoming light beam from the object to be measured.
- the filters may be of a fixed nature or tunable (AOTF or LCTF). In the latter case a single tunable filter is used to sequentially mimic the decorrelation matched filters under electronic control.
- the filters are matched to take advantage of the correlations between the spectral data derived from remote scenes for best results which include (i) increased signal to noise ratio due to averaging between the correlated data, and (ii) reduction of the amount of data and measurement time needed at the outset, due to the projection of the spectra onto a decorrelated parameter space.
- the number of filters required to achieve a good measurement is much lower than the number of wavelengths of the original spectral image so that the measurement itself is much shorter.
- FIG. 1 is a block diagram illustrating the main components of an imaging spectrometer constructed in accordance with U.S. Pat. No. 5,539,517 (prior art);
- FIG. 2 illustrates a Sagnac interferometer, as used in an imaging spectrometer in accordance with U.S. Pat. No. 5,539,517 (prior art);
- FIGs. 3a and 3b are chromosome images at spectral bands 724 nm and 597 nm, respectively, as derived from a spectral cube measured using the SPECTRACUBETM system;
- FIG. 4 is a graphic presentation of the twenty eigenvalues of the covariance matrix of the same spectral cube
- FIGs. 5a-e are the first five PCA images of the same spectral cube, shown after cube stretching to cover the full dynamic range of the display;
- FIG. 6 is a graphic presentation of matrix V shown in Table 4;
- FIG. 7 is a male color karyotype obtained using the first five PCs, the karyotype is arranged in chromosome pairs;
- FIG. 8a-c are three male color karyotypes, prior to arrangement of the chromosomes in pairs, obtained using the first three, four and five PCs, respectively;
- FIG. 9 is a graphic presentation of five decorrelation matched filters as calculated using the data in Tables 4 and 6 and Equation 6;
- FIG. 10 is a schematic depiction of a filter wheel including spectral decorrelation measurement apparatus suitable for chromosome analysis
- FIG. 11 is a schematic depiction of a tunable filter including spectral decorrelation measurement apparatus suitable for chromosome analysis
- FIG. 12 is a schematic depiction of a filter wheel including spectral decorrelation measurement apparatus suitable for remote sensing.
- the present invention is of a method and hardware for remote scenes classification by decorrelation statistical analysis which can be used to provide detailed information about the landscape in terms of composition, structure, properties, conditions, etc.
- the present invention can be used to design decorrelation matched optical filters, also referred herein as "matched" filters, for fast measurement aimed at remote scenes classification.
- scene refers to any element present in a field of view.
- a scene may be a corn field, another scene may be a stressed section of that or other field, yet another scene may be a point object, or as such objects are referred to in the art of remote sensing - a target.
- the term scene refers to anything which is scented via remote sensing.
- Specfral imaging is the technology that enables the measurement of the spectrum of electronic radiation reflected, emitted, or luminesced by every point (pixel) of a scene.
- a spectral imager also referred herein as imaging spectrometer
- a spectral image is a collection of spectra of the scenes measured by a spectral imager. It is usually organized as an intensity function defined in a three dimensional space in which two dimensions are of an image (x and y), and one is of a spectral axis ( ⁇ ).
- a spectral image is usually referred to as a "cube" of data or "spectral cube”.
- spectral cubes Prior art teaches different methods of measuring spectral images (i.e., spectral cubes).
- Devices designed according to these methods include light collection optics; a dispersion element (e.g., a grating), f ⁇ lter(s) (e.g., AOTF or LCTF) or an interferometer; focusing optics; and a two-dimensional array of detectors (typically a CCD in the visible range and other types of detectors in the infrared range, etc.).
- a dispersion element e.g., a grating
- f ⁇ lter(s) e.g., AOTF or LCTF
- interferometer e.g., a two-dimensional array of detectors
- detectors typically a CCD in the visible range and other types of detectors in the infrared range, etc.
- a specfral image is the next step after a color image.
- two fluorescent dyes such as Texas Red and Rhodamine appear the same color to the human eye but they are well distinguished by a spectrograph with ten nanometers resolution. Since the colors as perceived by the human eye are composed of combinations of only three colors, red, green and blue (RGB), the number of different regions in a scene that can be classified by color is very limited.
- a spectral imager For each point of the same scene a spectral imager measures a spectrum which depends on the chemical materials present at that point, and this is a function of wavelength which contains the order of fifty or one hundred data (depending on spectral resolution) instead of only three as for a color image.
- a spectral imager measures a spectrum which depends on the chemical materials present at that point, and this is a function of wavelength which contains the order of fifty or one hundred data (depending on spectral resolution) instead of only three as for a color image.
- the present invention is aimed at constructing a system which performs a measurement, with high sensitivity, and at high speed, and encompassing a small amount of data from the outset.
- the hardware according to the invention does not require an interferometer, but only a number (N) of what is herein referred to as "decorrelation matched filters" (fixed or tunable), which are designed using decorrelating statistical analysis such as principal component analysis (PCA) and are placed in the path of the incoming radiation from the scene to be measured.
- PCA principal component analysis
- Decorrelation enables to decrease the amount of data taking at the outset, while the correlations among the data are taken advantage of in order to increase the signal to noise ratio.
- 'decorrelation' refers to an algorithm which defines an initial set of correlated parameters into a new set of parameters which is the linear combination of the initial set, which new set of parameters are independent from each other and are then reduced to a minimal number which still carries the required information.
- a classification method based on the position of scenes in a multidimensional space defined by a decorrelated set of parameters is characterized by increased confidence level and shorter measurement time as compared with classification methods employing the initial set of parameters devoid of prior decorrelation.
- Each decorrelation matched filter is mathematically described by a weighting function whose shape is such that the parameters which are more correlated are added with a larger weight.
- Different filters in order to decorrelate the data, have a different shape and therefore weigh the initial parameters values in different ways.
- the signal obtained using each of the filters has contribution from all initial parameters, therefore the signal is measured with higher signal to noise ratio as compared with other, more conventional, methods.
- the decorrelation matched filters are dedicated for a given procedure in which the scenes share features with reference scenes used for obtaining a reference library for the decorrelation analysis.
- the decorrelation matched filters are placed in a filter wheel of the dedicated hardware according to the invention and are introduced successively in the radiation beam, while a suitable detector builds the images. That is to say that the detector builds an image with one filter, then the wheel rotates to another filter, and the detector builds a new image in synchronization, and so on until one image for each filter has been measured (N images).
- the decorrelation matched filters are mimicked by a tunable filter such as AOTF or LCTF which are tuned by a tuning device.
- FIG. 1 is a block diagram illustrating the main components of a prior art imaging spectrometer disclosed in U.S. Pat. No. 5,539,517, to Cabib et al., which is incorporated by reference as if fully set forth herein.
- This imaging spectrometer is constructed highly suitable to implement the method of the present invention as it has high specfral (Ca. 4-14 nm depending on wavelength) and spatial (Ca. 30/M ⁇ m where M is the effective fore optics magnification) resolutions.
- the prior art imaging spectrometer of Figure 1 includes: a collection optical system, generally designated 20; a one-dimensional scanner, as indicated by block 22; an optical path difference (OPD) generator or interferometer, as indicated by block 24; a one-dimensional or two-dimensional detector array, as indicated by block 26; and a signal processor and display, as indicated by block 28.
- a collection optical system generally designated 20
- a one-dimensional scanner as indicated by block 22
- OPD optical path difference
- interferometer as indicated by block 24
- a one-dimensional or two-dimensional detector array as indicated by block 26
- signal processor and display as indicated by block 28.
- a critical element in system 20 is the OPD generator or interferometer 24, which outputs modulated light corresponding to a predetermined set of linear combinations of the spectral intensity of the light emitted from each pixel of the scenes to be analyzed.
- the output of the interferometer is focused onto the detector array 26.
- the apparatus according to U.S. Pat. No. 5,539,517 may be practiced in a large variety of configurations. Specifically, the interferometer used may be combined with other mirrors as described in the relevant Figures of U.S. Pat. No. 5,539,517.
- interferometers may be employed. These include (1) a moving type interferometer in which the OPD is varied to modulate the light, namely, a Fabry-Perot interferometer with scanned thickness; (2) a Michelson type interferometer which includes a beamsplitter receiving the beam from an optical collection system and a scanner, and splitting the beam into two paths; (3) a Sagnac interferometer optionally combined with other optical means in which interferometer the OPD varies with the angle of incidence of the incoming radiation, such as the four- mirror plus beamsplitter interferometer as further described in the cited U.S. Pat. application (see Figure 14 there).
- a moving type interferometer in which the OPD is varied to modulate the light namely, a Fabry-Perot interferometer with scanned thickness
- Michelson type interferometer which includes a beamsplitter receiving the beam from an optical collection system and a scanner, and splitting the beam into two paths
- Sagnac interferometer optionally combined with other optical means in which
- Figure 2 illustrates an imaging spectrometer constructed in accordance with U.S. Pat. No. 5,539,517 utilizing an interferometer in which the OPD varies with the angle of incidence of the incoming radiation.
- a beam entering the interferometer at a small angle to the optical axis undergoes an OPD which varies substantially linearly with this angle.
- OPD's and therefore the spectrum of each pixel of the scene can be reconstructed by Fourier transformation.
- a beam parallel to the optical axis is compensated, and a beam at an angle (6) to the optical axis undergoes an OPD which is a function of the thickness of the beamsplitter 33, its index of refraction, and the angle ⁇ .
- the OPD is proportional to ⁇ for small angles.
- OPD0Q,r5, , «) ⁇ [( « 2 -sin 2 09+6)) 0 - 5 - ( « 2 -sin 2 0S-6)) 0 - 5 + 2sin ⁇ sinc>]
- ⁇ is the angle of incidence of the ray on the beamsplitter
- 6 is the angular distance of a ray from the optical axis or interferometer rotation angle with respect to the central position
- t is the thickness of the beamsplitter
- n is the index of refraction of the beamsplitter.
- Equation 1 it follows from Equation 1 that by scanning both positive and negative angles with respect to the central position, one can get a double-sided interferogram for every pixel, which helps eliminate phase errors, thereby giving more accurate results in the Fourier transform calculation.
- the scanning amplitude determines the maximum OPD reached, which is related to the spectral resolution of the measurement.
- the size of the angular steps determines the OPD step which is, in turn, dictated by the shortest wavelength to which the system is sensitive. In fact, according to the sampling theorem [see, Chamberlain (1979) The principles of interferometric spectroscopy, John Wiley and Sons, pp. 53-55], this OPD step must be smaller than half the shortest wavelength to which the system is sensitive.
- Another parameter which should be taken into account is the finite size of a detector element in the matrix.
- the element Through the focusing optics, the element subtends a finite OPD in the interferometer which has the effect of convolving the interferogram with a rectangular function. This brings about, as a consequence, a reduction of system sensitivity at short wavelengths, which drops to zero for wavelengths equal to or below the OPD subtended by the element. For this reason, one must ensure that the modulation transfer function (MTF) condition is satisfied, i.e., that the OPD subtended by a detector element in the interferometer must be smaller than the shortest wavelength at which the instrument is sensitive.
- MTF modulation transfer function
- 5,539,517 do not merely measure the intensity of light coming from every pixel in the field of view, but also measure the spectrum of each pixel in a predefined wavelength range. They also better utilize all the radiation associated with each pixel in the field of view at any given time, and therefore permit, as explained above, a significant decrease in the frame time and/or a significant increase in the sensitivity of the spectrometer.
- imaging spectrometers may include various types of interferometers and optical collection and focusing systems, and may therefore be used in a wide variety of applications, including remote sensing for geological, agricultural and military investigations, and the like.
- an imaging spectrometer in accordance with the invention disclosed in U.S. Pat. No. 5,539,517 was developed by Applied Spectral Imaging Ltd., Industrial Park, Migdal Haemek, Israel and is referred herein as SPECTRACUBETM.
- the SPECTRACUBETM system optically connected to a telescope is used to implement the method of the present invention.
- the SPECTRACUBETM system has the following characteristics, listed hereinbelow in Table 1 :
- Sensitivity 20 milliLux (for 100 msec integration time, increases for longer integration times linearly with r )
- the prior art SPECTRACUBETM system may be to acquire spectral data of every pixel of scenes in a field of view.
- any spectral imager i.e., an instrument that measures and stores in memory for later retrieval and analysis the spectrum of radiation associated with every point or pixel of scenes present in its field of view, including filter (e.g., acousto-optic tunable filters (AOTF) or liquid- crystal tunable filter (LCTF)) and dispersive element (e.g., grating) based spectral imagers can be used to acquire the required spectral data. Therefore, it is not intended to limit the scope of the present invention for use of any specific type of spectral imager.
- filter e.g., acousto-optic tunable filters (AOTF) or liquid- crystal tunable filter (LCTF)
- dispersive element e.g., grating
- the wavelength range of choice will determine to some extent the exact features of the system. For example, should an interferometric based spectral imager be the choice, then for the spectral range of 300-400 nm the optical components, such as the beam splitter, the optics and the array detector should be made of materials suitable for UV and coated with special UV enhanced coatings; for the spectral range of 400-1,000 nm the optical components, such as the beam splitter, the optics an the array detector should be made of materials suitable for visible radiation and coated with special coatings which optimize the performance in the visible range. The mirrors of the interferometer should be coated for maximum reflection in the various spectral ranges with suitable coatings.
- the array detectors are special detectors based not on Silicon technology but on InSb (Indium Antimonide), HgCdTe (Mercury Cadmium Telluride), and others.
- InSb Indium Antimonide
- HgCdTe Mercury Cadmium Telluride
- a method and hardware for remote scenes classification by decorrelation statistical analysis which can be used to provide information about the landscape and targets thereon.
- a method for preparing a reference template for classification of remote scenes includes the steps of (a) classifying a set of reference scenes via a conventional classification technique for obtaining a set of preclassified scenes; a reference scene is a scene of known nature, e.g., soil composition; (b) using a specfral imager for measuring a spectral cube of the preclassified referenced scenes; (c) employing a decorrelation statistical method for exfracting the spectral cube for decorrelated specfral data characterizing the reference scenes; and (d) using at least a part of the decorrelated specfral data for the preparation of the reference template for remote scenes classification.
- the spectral imager may be of any type.
- the spectral imager may include an element selected from the group consisting of a dispersion element, a filter, a tunable filter and an interferometer.
- the decorrelation statistical method may be of any type, including, but not limited to, principal component analysis, canonical variable analysis and/or singular value decomposition.
- the principal component analysis includes expressing each of the scenes as linear combinations of N eigenvectors.
- N is an integer greater than two. More preferably, N is an integer greater than two and smaller than eight.
- the principal component analysis includes the steps of (a) selecting k spectral slices for the spectral cube of the reference scenes; (b) calculating an average spectrum for each of the reference scenes; (c) stretching each of the average spectra for obtaining a stretched average spectrum for each of the reference scenes; (d) averaging the stretched average spectra for each of the reference scenes, for obtaining an ensemble average spectrum for each of the reference scenes; (e) calculating a k dimension eigen system for the ensemble average spectra and exfracting N eigenvectors; (f) using the N eigenvectors for defining an N-dimension vector for each of the reference scenes; and (g) using the N-dimension vectors for preparing the reference template for the remote scenes classification, k is preferably an integer greater than two, preferably greater than nine.
- the principal component analysis further includes the step of (h) performing a spatial averaging procedure over all specfral slices.
- the method includes the steps of (a) preparing a reference template for classification of the remote scenes via (i) classifying a set of reference scenes via a conventional classification technique for obtaining a set of preclassified reference scenes; (ii) using a first spectral imager for measuring a spectral cube of the preclassified reference scenes; (iii) employing a principal component analysis for exfracting the specfral cube for decorrelated specfral data characterizing the reference scenes; and (vi) using at least a part of the decorrelated specfral data for the preparation of the reference template for remote scenes classification; (b) using a second spectral imager for measuring a specfral cube of analyzed remote scenes, such that a spectrum of each pixel in the remote scenes is obtained; (c) employing a decorrelation statistical method for exfracting decorrelated spectral data characterizing the pixels; and (d) comparing at least a part of the decorrelated specfral data extracted from
- the method includes the steps of (a) preparing a reference template for classification of remote scenes via (i) classifying a set of reference scenes via a conventional classification technique for obtaining a set of preclassified reference scenes; (ii) using a first specfral imager for measuring a specfral cube of the preclassified reference scenes; (iii) employing a decorrelation statistical method for extracting the specfral cube for decorrelated spectral data characterizing the reference scenes; and (iv) using at least a part of the decorrelated specfral data for the preparation of the reference template for the remote scenes classification; (b) using a second spectral imager for measuring a spectral cube of analyzed remote scenes, such that a spectrum of each pixel in the remote scenes is obtained; (c) projecting the spectrum of each of the pixels onto the decorrelated specfral data for obtaining a projected spectrum for each of the pixels; and (d) comparing the projected specfra with
- the method includes the steps of (a) preparing a reference template for classification of remote scenes via (i) classifying a set of reference scenes via a conventional classification technique for obtaining a set of preclassified reference scenes; (ii) using a first specfral imager for measuring a spectral cube of the preclassified reference scenes; (iii) employing a principal component analysis for extracting the specfral cube for decorrelated spectral data characterizing the reference scenes via (a) selecting k specfral slices for the specfral cube of the reference scenes; (b) calculating an average spectrum for each of the reference scenes; (c) stretching the average spectra for obtaining a stretched average spectrum for each of the reference scenes; (d) averaging the stretched average spectra for each of the reference scenes for obtaining an ensemble average spectrum for each of the reference scenes; (e) calculating a k dimension eigen system for the ensemble average specfra and extracting N eigenve
- the method further includes the step of performing a spatial averaging procedure on all specfral slices.
- Fifth provided is a method of calculating decorrelation matched filters for remote scenes classification, the decorrelation matched filters being for extracting decorrelated spectral data from the remote scenes, the method comprising the step of (a) obtaining decorrelated spectral data characterizing a set of reference scenes via (i) classifying the set of reference scenes via a conventional classification technique for obtaining a set of preclassified reference scenes; (ii) using a first spectral imager for measuring a spectral cube of the preclassified reference scenes; and (iii) employing a decorrelation statistical method for extracting the spectral cube for decorrelated specfral data characterizing the reference scenes; and (b) mathematically manipulating at least a part of the decorrelated specfral data for obtaining a mathematical description of the decorrelation matched filters.
- the decorrelated specfral data is obtained using a principal component analysis, which includes expressing each of the reference scenes by a linear combination of N eigenvectors.
- a set of decorrelation matched filters for remote scenes classification the decorrelation matched filters being for extracting decorrelated spectral data from the remote scenes
- the set includes physical filters having shapes, the shapes following a mathematical description, the mathematical description being obtainable by (a) obtaining decorrelated spectral data characterizing a set of reference scenes via (i) classifying the set of reference scenes via a conventional classification technique for obtaining a set of preclassified reference scenes; (ii) using a first specfral imager for measuring a specfral cube of the preclassified reference scenes; and (iii) employing a decorrelation statistical method for exfracting the specfral cube for decorrelated spectral data characterizing the reference scenes; and (b) mathematically manipulating at least a part of the decorrelated spectral data for obtaining the mathematical description of the decorrelation matched filters.
- a method of tuning a tunable filter for remote scenes classification the method renders the tunable filter to mimic a set of decorrelation matched filters, and is for extracting decorrelated specfral data from the remote scenes
- the method includes the steps of (a) obtaining decorrelated specfral data characterizing a set of reference scenes via (i) classifying the set of reference scenes via a conventional classification technique for obtaining a set of preclassified reference scenes; (ii) using a first spectral imager for measuring a specfral cube of the preclassified reference scenes; and (iii) employing a decorrelation statistical method for exfracting the specfral cube for decorrelated spectral data characterizing the reference scenes; (b) mathematically manipulating at least a part of the decorrelated specfral data for obtaining a mathematical description describing the set of decorrelation matched filters; and (c) sequentially tuning the tunable filter according to the mathematical description.
- the t decorrelated specfral data characterizing
- the method includes the steps of (a) preparing a reference template for classification of the remote scenes via (i) classifying a set of reference scenes via a conventional classification technique for obtaining a set of preclassified reference scenes; (ii) using a first specfral imager for measuring a specfral cube of the preclassified reference scenes; (iii) employing a decorrelation statistical method for exfracting the spectral cube for decorrelated spectral data characterizing the reference scenes; and (iv) using at least a part of the decorrelated specfral data for the preparation of the reference template for the remote scenes classification; (b) calculating a mathematical description of decorrelation matched filters for classification of the remote scenes employing the reference template, the calculation being by mathematically manipulating at least a part of the decorrelated spectral data; (c) using the mathematical description of the decorrelation matched filters for manufacturing the decorrelation matched filters; (d) using the decorrelation matched filters for ex
- the method includes the steps of (a) providing a set of decorrelation matched filters for the remote scenes classification, the decorrelation matched filters being for extracting decorrelated spectral data from the remote scenes, the set including physical filters having shapes, the shapes following a mathematical description, the mathematical description being achieved by (i) preparing a reference template for classification of remote scenes via (a) classifying a set of reference scenes via a conventional classification technique for obtaining a set of preclassified reference scenes; (b) using a first spectral imager for measuring a specfral cube of the preclassified reference scenes; (c) employing a decorrelation statistical method for extracting the specfral cube for decorrelated specfral data characterizing the reference scenes; and (d) using at least a part of the decorrelated spectral data for the preparation of the reference template for the remote scenes classification; and (ii) mathematically manipulating at least a part of the decorrelated spectral data for
- the method further includes the step of (d) according to the comparison, attributing each pixel an artificial color.
- the method includes the steps of (a) providing a tunable filter and tuning information for tuning the tunable filter so as to mimic a set of decorrelation matched filters, the tunable filter being for extracting decorrelated spectral data from the remote scenes, the tuning information being achieved by (i) preparing a reference template for classification of remote scenes via (a) classifying a set of reference scenes via a conventional classification technique for obtaining a set of preclassified reference scenes; (b) using a first spectral imager for measuring a specfral cube of the preclassified reference scenes; (c) employing a decorrelation statistical method for exfracting the specfral cube for decorrelated specfral data characterizing the reference scenes; and (d) using at least a part of the decorrelated specfral data for the preparation of the reference template for the remote scenes
- a specfral decorrelation measurement apparatus for remote scenes classification by exfracting decorrelated specfral data from the remote scenes
- the apparatus is connected to a telescope used to view the remote scenes
- the apparatus includes (a) a detector; and (b) an optical system for transmitting electromagnetic radiation from the remote scenes onto the detector, the optical system including a set of decorrelating matched filters, the decorrelation matched filters being for extracting decorrelated spectral data from the remote scenes, the filters of the set of decorrelation matched filters having shapes, the shapes following a mathematical description, the mathematical description being calculated by (i) obtaining decorrelated specfral data characterizing a set of reference scenes via (a) classifying the set of reference scenes via a conventional classification technique for obtaining a set of preclassified reference scenes; (b) using a first spectral imager for measuring a spectral cube of the preclassified reference scenes; and (c) employing a decorrelation statistical method for extracting the spectral cube for de
- the optical system further includes a collimating lens for collimating radiation reaching any of the decorrelating matched filters and focusing optics to image the scene on an array detector.
- the decorrelation matched filters are arranged on a rotatable filter carrying element.
- specfral decorrelation measurement apparatus for remote scenes classification by extracting decorrelated specfral data from the remote scenes
- the apparatus is connected to a telescope used to view the remote scenes
- the apparatus includes (a) a detector; and (b) an optical system for transmitting electromagnetic radiation from the remote scenes onto the detector, the optical system including a tunable filter and a tuning device, the tuning device being for tuning the tunable filter, so that the tunable filter sequentially mimics a set of decorrelating matched filters, the decorrelation matched filters mimicked by the tunable filter being for exfracting decorrelated spectral data from the remote scenes, the tuning of the tunable filter being calculated according to a mathematical description, the mathematical description being calculated by (i) obtaining decorrelated spectral data characterizing a set of reference scenes via (a) classifying the set of reference scenes via a conventional classification technique for obtaining a set of preclassified reference scenes; (b) using a first spectral imager for measuring a specfral cube
- chromosome samples are very small and close scenes viewed via a microscope.
- the principles of the analysis are identical, since once a specfra cube is measured, the decorrelation statistical analysis that follows is substantially identical.
- spectral imagers which suit both remote as well as close scenes measurements are well known.
- One example is the SPECTRACUBETM system described hereinabove which may be connected to various optical instruments including microscopes and telescopes.
- spectral bio-imaging which is a combination of Fourier spectroscopy, CCD-imaging and optical microscopy enabling the measurement of accurate specfral data simultaneously at all points of a biological sample, was used to visualize hybridization based multicolor appearance of all (i.e., 24) types of human chromosomes and to generate a color map of the human karyotype.
- data spectrally collected by any spectral imager having a specfral resolution of 14 nm or higher is statistically analyzed using decorrelation statistical methods, such as, but not limited to, principal component analysis (PCA) to find correlations among the data and to construct (i) a reference template which may then be used for routine analysis of new samples; and (ii) a hardware based on decorrelation matched filters (fixed or tunable) for collecting only the decorrelated spectral data (a smaller amount than with prior art), from new samples while the correlated data are averaged over at the outset, thereby decreasing measurement time and increasing signal to noise ratio.
- PCA principal component analysis
- Both the reference template and the decorrelation matched filters can be used to classify chromosomes and to detect chromosomal aberrations. Similarly, they can be used to analyze remote scenes.
- both the reference template and the decorrelation matched filters are dedicated to a given experimental procedure in which the dyes and their combinations which are used to label any of the chromosomes are predetermined.
- 24 chromosome paints (1 through 22, X and Y, Table 2), each labeled with a different combination of three or less different flourophores selected from a set of five fluorophores according to the combinatorial hybridization approach (a through e, Table 2), (see Table 2 for the different fluorophores and their spectral characteristics and Table 3 for the assignment of the fluorophores listed in Table 2 to obtain the 24 chromosome paints), were simultaneously hybridized with human mitotic chromosome spreads of few non-related male white blood cells, prepared for hybridization essentially as described in Ried et al.
- Hybridized chromosomes were viewed through an inverted fluorescence microscope connected to the SPECTRACUBETM System and were analyzed.
- Decorrelation statistical analysis is directed at exfracting decorrelated data out of a greater amount of data, and average over the correlated portions thereof.
- PCA principal component analysis
- canonical variable analysis and singular value decomposition, etc. of these methods PCA is perhaps the more common one, and its use for decorrelation of spectral data, as this term is defined above, is hereinafter described.
- Principal component analysis is one of a number of powerful techniques used in multivariate statistical analysis. It is advantageous in cases where a large number of "results", which depend on a large number of possibly correlated variables forms the basic data set. Its strength lies in the fact that this data decomposition provides a transformation to decorrelated variables, while simultaneously averaging over correlated variables.
- the basic data set i.e., the specfral cube
- the basic data set is composed of k specfral slices of the same field, where each spectral slice is obtained at a different specfral band.
- the data set is composed of the specfra of all the pixels of the field.
- One of the objectives of looking at such a data set can be the characterization of the pixels into groups of similar specfra.
- each spectral slice as a vector whose elements are the image pixels arranged into the column vector using a predetermined code.
- the matrix x ⁇ x nm ⁇ carries the full information, so that each column is a spectral slice.
- a matrix y derived from matrix x by subtracting from each column, the column average.
- the various columns of the y matrix may be correlated, so that, some of the information carried by the data is correlated.
- the PCA technique decorrelates the information and reduces it only to decorrelated variables, so that the amount of "real" data pixels is smaller and easier to handle.
- y' is the transpose of y.
- the i,j term of c is the covariance of the i-th slice with the j-th slice, i.e. if they are decorrelated this term vanishes.
- the diagonal of c is composed of the variances of each specfral slice, which can be regarded as a scale for the amount of information in this particular slice. Alternatively, this variance (its square root) can be regarded as the average contrast of this particular slice. Linear algebra describes this situation as follows. The elements of interest
- the eigenvalues vanish.
- the number of non- vanishing eigen- values defines the dimension of the information, which dimension is smaller than k.
- the corresponding eigen-vectors define a subspace in the original k space in which the full information content is represented. Furthermore, the information in each new dimension is completely decorrelated to the information in the other dimensions. Thus in the new space the full information content is represented in a decorrelated manner so that it can be easily used for classification purposes.
- such an analysis can be performed on a whole spectral cube.
- the analysis is performed only for selected pixels or mathematically manipulated (e.g., after background or interferences subtraction and averaging) selected pixels to improve the results and enable better classification later on.
- the preferred approach is described in more detail below, nevertheless, there is no intention to limit the scope of the present invention to the preferred approach employed, as different mathematical manipulations may be found useful for different data collection approaches (e.g., filter or dispersion element based specfral imagers) and/or different radiation employed.
- the results of the specfral measurements of the stained chromosomes sample are stored in a spectral cube in any convenient format.
- a conversion is typically needed. The conversion can be performed by a dedicated software able to read each pixel in each spectral slice and write it into another file in a format suitable to the user.
- a commercial software sold under the name ENVI (environment for visualizing images) by Research Systems Inc., Boulder Co., USA, can be used.
- ENVI can read a specfral cube and write, on request, each specfral slice into a standard *.GIF format which can be read by a large number of routines.
- ENVI can be used in a different mode called the BSQ format. The later writes the cube in a binary sequential manner, starting at the spectral slice possessing the highest wave number, wherein each slice is written column after column.
- Matlab a popular mathematical analysis software and programming environment, with the addition of special software packages able to read the spectral cubes.
- a dedicated software package can be written in an appropriate language, either stand alone or incorporated in the existing SPECTRACUBETM system software, to perform the building of the basic data set.
- specfral slices were those where significant contrast and low noise was present. It should be noted that any other number of specfral slices from about 10-15 or more (up to a limit which is determined by the spectral resolution of the spectral imager and the spectral range) can be used and that there is no intention to limit the scope of the invention to any specific number.
- Figures 3a-b present two examples of the spectral slices, at 724 nm ( Figure 3 a) and 597 nm ( Figure 3b). Notice the difference in contrast, noise and intensity over the various chromosomes.
- the k (e.g., twenty) specfral slices above are first used to find the decorrelated (and therefore orthogonal) principal components vectors of each chromosome type (L types in a species, 24 types in human), then the projection of each chromosome in each relevant principal component direction is calculated . This is equivalent to find the position of each chromosome template in the principal component space. This is performed by first identifying the chromosomes via conventional chromosome classification techniques, e.g., G- banding, R-banding or color karyotyping as described in E. Schroeck et al. (1996) Multicolor specfral karyotyping of human chromosomes, Science, 273, 494-497. It should be noted that the method described in Science magazine can be performed post G-banding or simultaneously to R-banding without affecting the measured specfral results.
- conventional chromosome classification techniques e.g., G- banding, R-banding or color karyotyping as
- each of the chromosomes is labeled with a different fluorophore or a different combination of fluorophores
- the projection of each chromosome onto each of the orthogonal principal components (PCs) forms a specifying reference vector (24 different vectors for a human male, 23 for a human female, different numbers for other species, generally L vectors for a species) which is unique to each of the chromosomes.
- L reference vectors collectively form a reference template for chromosome classification, as each of these L reference vectors is used as an identification means to attribute each pixel of a new specfral cube to one of the L chromosome types.
- these L different reference vectors i.e., the reference template
- This is done by a classification algorithm which compares measured vectors of pixels in a cube to be analyzed to the above reference vectors, followed by attributing each of the pixels a matching artificial color selected from L different artificial colors, according to the comparison results.
- a preferred embodiment for pre-processing for classification is herein described in more detail:
- each staining technique at least one, preferably a number, of independently classified (e.g., G-banding) images are selected, from which a reference template is calculated as follows.
- a 3 x 3 spatial averaging is performed on all selected k spectral slices such that each nine pixels are now attributed an average spectrum and are considered a single enlarged pixel.
- this procedure reduces the spatial resolution of the analysis, it is employed as it also reduces noise to a large extent. It should however be noted that this step was experimentally determined to be useful, yet as the amount of noise associated with other data collection approaches and/or experimental procedures may differ tremendously, there is no intention to limit the invention to a specific mode of averaging (e.g., 3 x 3) or to averaging altogether.
- an average spectrum of the whole cube(s), or the average spectrum of some, most or all the background pixels is calculated and is thereafter subfracted from the spectrum of each of the pixels in the cube. It will be appreciated that as most of the pixels in a cube are of background, the different calculated average spectra yield nearly identical results. Nevertheless, it should be noted that as the intensity of background may differ to a large extent while using other data collection approaches and/or experimental procedures, different background subtraction procedures may be employed. Furthermore, as in some cases background subtraction may be found improving results in a marginal way only, it may therefore, in these cases, be discarded altogether.
- each chromosome is masked, typically manually or with a suitable automatic image processing algorithm, and the average spectrum for each chromosome within its mask is calculated.
- the analysis can be performed on a whole spectral cube(s), yet preferably, the analysis is performed only for selected pixels or mathematically manipulated (e.g., after background subtraction and averaging) selected pixels to improve the results and enable better classification later-on.
- the average spectrum of each chromosome calculated above is separately stretched preferably into the maximal dynamic range 0-1.
- L types, 24 types for human are the chromosome types derived from different or same cube measurements.
- the PCA will be performed on mathematically manipulated pixels. These are L (e.g., 24 for human) pixels each having a unique ensemble average spectrum which are mathematically manipulated as described. It should be noted, however, that one can obtain similar results also when applying the above described steps in a different order.
- Sixth, an eigen system of the ensemble average spectra is calculated and the eigenvectors are used to calculate the PC of the ensemble average specfra.
- the L chromosomes ensemble average specfra are stored in an L (e.g., 24 for human) line by k (e.g., 20) columns matrix, called matrix S, where each line represents a chromosome and each column a given specfral slice which are selected as described above under Example 2.
- L e.g., 24 for human
- k e.g., 20
- matrix S matrix
- each line represents a chromosome and each column a given specfral slice which are selected as described above under Example 2.
- the number k of spectral slices is not necessarily 20 and can be any number between substantially 10-15 or more, say 40, about equally spaced in the 500-800 nm region, depending on the emission spectra of the fluorophores employed, k is preferably selected to be an integer greater than nine.
- C is defined as the covariance of S.
- C is typically a k x k (e.g., 20 x 20) matrix which represents the covariance between each pair of chromosomes.
- the eigen system of matrix C is calculated.
- V is defined as the matrix composed of the eigenvectors (each column being an eigenvector).
- Figure 4 presents 20 eigenvalues corresponding to these eigenvectors in decreasing order. Please note that this is a logarithmic display. From the graph presented in Figure 4, one notices that the eigenvalues decrease very sharply with increasing number. From the point of view of the information content of the specfral cube, one can say that more than 95 % of the information is contained in the first PCA image. On observing the PCA images, one notices that the chromosomes do appear only in the first 7 components, the first five of which are shown in Figures 5a-e.
- Table 4 presents the first five (most significant) eigenvectors (Vec 1 - Vec 5) for this system. Vec 1 being the most significant, whereas the first 20 lines of each column represent the relative weight of the wavelength depicted in the leftmost column.
- Figure 6 is a graphic presentation of the first five vectors of matrix V. TABLE 4
- SS is a 24 by 20 matrix.
- N eigenvalues are of significance, typically 3-5, thus for chromosomes, typically N is selected to be an integer greater than two. Consequently only about N (e.g., 3 to 5) out of the k (e.g., 20) numbers composing a line of SS in the given example are of any significance, additional eigenvectors and eigenvalues are of marginal significance and may therefore be discarded.
- N e.g., 3 to 5
- a chromosome is uniquely specified by its N (e.g., 5) coordinates in the N-dimensional space with respect to the N chosen eigenvectors (and eigenvalues). These N coordinates represent the coefficients of the linear combination of the chromosome in question with respect to the N orthogonal eigenvectors.
- N e.g. 5
- Using this type of analysis as compared with forming the matrix S for all the pixels in the cube ensures that (i) background pixels are not introduced into the calculation, and therefore the noise associated with them is avoided, and that (ii) each of the chromosomes is equally represented in terms of number of pixels.
- Table 5 presents an example of a truncated SS matrix. In this Table each row represents the projection of a given human chromosome on the five orthogonal eigenvectors or principal components (PCs) describing the spectral cube(s).
- the vectors of Table 5 form a reference template which may thereafter be used for classification of all the 24 chromosomes of a newly and similarly measured cube, as described hereinbelow in the following Examples.
- Each of these L (24 in the given example) N-dimension vectors (five- dimension vectors in the given example), P, is checked, by calculating the correlation, against the L template vectors, in order to decide to which chromosome type it belongs.
- L N-dimension vectors may then serve as attraction centers for classification as delineated hereinbelow.
- 'attraction center' refers to a statistical method capable of attributing a pixel to a chromosome.
- Examples include but are not limited to (i) a minimum "distance" calculation in the
- N-dimensional space according to Equation 5, in which a "distance” is defined (for example the Euclidean definition or square root of the sum of the squares of the coordinate differences) and then the chromosome which is "closer” according to this definition is chosen, or (ii) a maximum correlation calculation in which the scalar products between the pixel vector and the L template vectors are calculated, and then the one that gives the highest result is chosen.
- a distance for example the Euclidean definition or square root of the sum of the squares of the coordinate differences
- the spectrum of each pixel in the cube is projected onto the average specfra principal domain, which is, in the preferred embodiment the 10- 15 or more, e.g., 40, preferably 20 dimension space of wavelengths or in other words the k spectral slices, to obtain an N dimension vector for each pixel.
- projection of the spectrum of each of the pixels is via the scalar product of the spectrum vector with the orthogonal base vectors of the PC space.
- pixels attributed to each of the L (24 in the given example) different chromosome types are given a different artificial color to yield an L (24) color karyotype.
- Figure 7 presents such a color karyotype, wherein the 24 chromosome types are arranged in pairs. Further presented in Figure 7 is a color key, according to which the chromosomes can be identified.
- Figures 8a-c present color karyotyping results obtained using 24 N- dimension vectors, wherein N equals 3, 4 and 5, respectively. Please note that best results are achieved using 24 5-dimension vectors (Figure 8c), yet, good results are obtained also with 24 4-dimension vectors ( Figure 8b). Furthermore, please note that using 24 3-dimension vectors as in Figure 8a enables unambiguous color classification of most chromosome pairs. These results are not surprising considering the above mentioned fact that more than 95 % of the information is contained in the first PCA image.
- each chromosome possesses a specific specfra.
- chromosome classification is performed by matching the spectra to a predetermined template which is derived using for example PCA.
- N is an integer greater than two, e.g., three or five, for each pixel of the image, can be measured directly, using imaging microscopy and N pre-designed filters.
- Equation 6 defines a physical filter F (representing transmission in the range 0-1):
- Vj k mm equals minimum (Vjfc) over all i
- max equals maximum (Vjt) over all i.
- N such filters (one for each k), wherein N is an integer greater than two, e.g., three or five. These N filters are physically realizable.
- each pixel in the image can be measured to produce an N-dimension vector for each.
- V vectors All the components of the V vectors are known from the stage of reference template preparation described hereinabove under Example 4.
- Table 6 represents the minimum and maximum value of the eigenvectors shown in Table 4 above.
- the sum ⁇ M j is the intensity of the pixel under discussion, which is measured directly from the microscope (no filter used).
- Figure 9 graphically depicts five decorrelation matched filters as calculated using the data in tables 4 and 6 and Equation 6.
- each of the PCs shown in the graph of Figure 6 has at least one positive section and at least one negative section
- each of these PCs may be represented by two or more physical filters in which one or some filters represents the positive section(s) of the PC and the other(s) represent the positive magnitude of the negative section(s) of the same PC. It should be further noted that some PCs may have only positive or only negative values.
- they may be represented accordingly by one or more filters which represent the positive values of the PC, or, one or more filters which represent the positive magnitude of the negative values of the PC, respectively.
- a measurement using these filters can be performed and the P N-dimension vector for each pixel can be determined and used for classification similarly to as described above.
- any of the filters presented in Figure 9, or any other filters differently calculated, for example as described above, may be manufactured as a single filter, alternatively as a subset of few filters, which collectively, when sequentially applied for measurement, yield otherwise substantially identical results. It should thus be noted that when the term decorrelation matched filters is used herein and especially in the claims section below, it refers to all the possible options of calculating and manufacturing these filters, unless otherwise is specifically indicated.
- Tunable filters such as acousto-optic tunable filters (AOTFs) and liquid-crystal tunable filters (LCTFs)
- TFs are solid state electronically tunable spectral bandpass selectors having no moving parts which can be electronically tuned to any particular wavelength, as well known in the art.
- a tunable filter can be thought of a variable bandpass filter that can be electronically tuned to any wavelength over its range.
- a liquid-crystal tunable filter is a solid state electronically tunable spectral bandpass filter typically made of high molecular weight organic substances having a dipole. Tuning LCTF is performed by subjecting the liquid crystal to varying electrical voltages.
- LCTF is a birefringent filter that uses phase retardation to create constructive and destructive interference. By stacking a number of stages in series, a single bandpass is obtained in a manner similar to that of a multicavity interference filter.
- LCTF technology was introduced by Cambridge Research & Instrumentation (CRI) Inc. in 1992.
- the first generation LCTFs produced suffered various limitations as far as bandpass width and shape and transmission of polarized and especially of randomly-polarized light are concerned.
- second generation LCTFs have overcome these problems, enabling transmission of about 100 percent of polarized light, substantially greater than 50 percent of randomly-polarized light, broad bandpass (top and bottom) of variety of shapes in the specfral range of 400 nm to 720 nm.
- the physical filters of Figure 9 can be mimicked by a single LCTF, which can be tuned at different times to mimic a filter of any bandpass of any desirable shape.
- An acousto-optic tunable filter is a solid state electronically tunable spectral bandpass filter which can be operated from the ultra violet through the visible and into the infrared regions of the optical spectrum.
- the AOTF operates on the principle of acousto-optic interaction in an anisotropic medium.
- the AOTF functions by the interaction of light with traveling acoustic wave through the medium, which creates a periodic modulation of its index of refraction by means of the elasto-optic effect. This modulation acts as a three-dimensional sinusoidal phase grating for light incident upon the crystal, leading to the diffraction of certain wavelengths at an angle from the incident beam radiation.
- an acoustic transducer typically a piezoelectric motor, is bonded to one face of the crystal and an acousto absorber is typically bonded to an opposite face.
- the transducer converts a high frequency rf (radio frequency) signal into a sinusoidal pressure wave which propagates laterally through the crystal.
- rf radio frequency
- the medium operates similar to a grating, wherein incident light is diffracted to its spectral wavelengths, light of varying wavelengths is acquired different angles with respect to the incident light beam when leaving the medium as a throughput.
- the acoustic absorber at the opposite end of the crystal eliminates acoustic reflections which would corrupt the primary acoustic wave form.
- AOTFs were used to generate a varying narrow bandpass, Nevertheless, electronically controlling the acousto wave parameters by for example super imposition (e.g., linear combination) acoustic waves of different wavelengths and/or different amplitudes, by for example employing more than one
- the AOTF when driven with multiple closely spaced rf s, the AOTF also provides electronically variable bandpass and shape control. To this effect the reader is referred to Eliot et al. (1996) Imaging acousto-optic tunable filter with 0.35 -micrometer spatial resolution. Applied Optics, 35, 5220-5226.
- any of the filters presented in Figure 9, or any other filters differently calculated, for example as described above in Example 6, may be manufactured as a single filter, alternatively as a subset of few filters, which
- any such combination of filters may be mimicked by a single tunable filter (LCTF or AOTF) which can be tuned at a different bandpass and shape to sequentially mimic any of the filters.
- LCTF tunable filter
- AOTF AOTF
- tunable filters such as AOTF and LCTF
- a single filter is required for measurement, the tunable filter is tuned to change its specfral characteristics in a manner that sequentially follows any desired characteristics.
- tuning information is selected such that the tunable filter sequentially mimics decorrelation matched filters. This, 5 however implies that the measurement involves no moving parts as it is electronically controlled.
- EXAMPLE 8 A Spectral Decorrelation Measurement Apparatus Based on Decorrelation l o Matched Optical Filters for Chromosome Analysis
- N e.g., three to five
- Apparatus 100 is 5 connected to a microscope 101, which is indicated by its objective lens 102.
- a sample of in situ painted chromosomes 104 to be analyzed is placed under microscope 101, on a supporting plane 106.
- Apparatus 100 further includes an optical system 108, which is for transmitting excitation light from light source 109 to sample 104 and emission light from sample 104 onto a detector 110, typically a
- optical system 108 includes an excitation filter 112, which is placed in the path of light emitted from light source 109.
- Excitation filter 112 is capable of transmitting light in the range required for excitation of the fluorophores in sample 104, e.g., in the ultraviolet and blue ranges, and of
- Optical system 108 further includes a dichroic filter 114, typically a triple dichroic filter, for directing exiting light from filter 112 to sample 104 and emission light from sample 104 to detector 110.
- optical system 108 further includes a barrier filter 116 for blocking any residual photons which are not in the specfral range of the emitted fluorescence.
- optical system 108 may further include a collimating lens 118 to ensure full collimation of the light. However, as well known in the art, some microscopes include a collimating lens themselves. In these cases collimating lens 118 may be discarded.
- Optical system 108 further includes N (N is an integer greater than two, preferably N is in the range of 3-5) decorrelating matched filters 120, five of which are shown in Figure 10, peripherally arranged on a rotatable filter carrying element 122, such as a filter wheel.
- N is an integer greater than two, preferably N is in the range of 3-5
- Each of decorrelating matched filters 120 is designed as described hereinabove under Example 6.
- Rotatable filter carrying element 122 also includes one position 124 through which light passes undisturbed.
- the number N of decorrelating matched filters 120 may vary as described above and is determined by the number of eigenvalues or PCs employed to construct the reference vector for each of the chromosomes, or, in other words, the reference template.
- Optical system 108 further includes a focusing lens 126 for focusing light after passage through rotatable filter carrying element 122 onto detector 110.
- apparatus 100 The operation of apparatus 100 is as follows. Decorrelating matched filters 120 of rotatable filter carrying element 122 are kept successively in the light beam while detector 110 builds an images for each. That is to say that detector 110 builds an image with first filter 120, then rotatable filter carrying element 122 rotates to present another filter 120, and detector 110 starts building a new image in synchronization, and so on until one image for each filter 120 has been measured. One additional image is formed while position 124 through which light passes undisturbed is positioned in the path of light.
- Equation 8 the coordinates of the P N-dimension vector for each pixel are calculated in the N-dimensional space of the N PCs. These coordinates are then used for classification by employing the pre-prepared reference template, similarly to as described above under classification. As before, for color presentation, pixels attributed to each of the 24 different chromosome types are given a different artificial color to yield a 24 color karyotype.
- a set of N e.g., three to decorrelation matched filters can be calculated and implemented by electronically tuning a tunable filter such as an AOTF or LCTF. Any of these filters can be used for fast collection of decorrelated projection of each pixel spectrum in a field of view onto a number of orthogonal PCs, provided that the observed object is treated according to the experimental protocol employed for calculating the fransmittance function of the filters as implemented by tuning.
- These N values collectively form an N-dimension vector for each pixel.
- Each of these vectors is then compared to reference vectors forming a reference template for classification as described above, and, based on this comparison, each pixel is attributed to a chromosome type and for presentation given a specifying artificial color.
- a tunable filter to serve as the decorrelation matched filters is placed in an apparatus which is referred to hereinbelow as a spectral decorrelation measurement apparatus, or apparatus 100'.
- Apparatus 100' is connected to a microscope 101', which is indicated by its objective lens 102'.
- a sample of in situ painted chromosomes 104' to be analyzed is placed under microscope 101', on a supporting plane 106'.
- Apparatus 100' further includes an optical system 108', which is for transmitting excitation light from light source 109' to sample 104' and emission light from sample 104' onto a detector 110', typically a two dimensional CCD array.
- optical system 108' includes an excitation filter
- Excitation filter 112' is capable of transmitting light in the range required for excitation of the fluorophores in sample 104', e.g., in the ultraviolet and blue ranges, and of blocking light in the range of fluorescent emission.
- Optical system 108' further includes a dichroic filter 114', typically a triple dichroic filter, for directing exiting light from filter 112' to sample 104' and emission light from sample 104' to detector 110'.
- optical system 108' further includes a barrier filter 116' for blocking any residual photons which are not in the specfral range of the emitted fluorescence.
- optical system 108' may further include a collimating lens 118' to ensure full collimation of the light.
- collimating lens 118' may be discarded.
- Optical system 108' further includes a tunable filter 120' and a tuning device 122'.
- Tuning device 122' is for sequentially tuning filter 120' according to precalculated tuning information to sequentially mimic N (N is an integer greater than two, preferably N is in the range of 3-5) decorrelating matched filters as described above under Example 7.
- Device 122' preferably also includes tuning information to transform filter 120' into a transparent optical element through which light passes undisturbed.
- the number N of mimicked decorrelating matched filters may vary as described above and is determined by the number of eigenvalues or PCs employed to construct the reference vector for each of the chromosomes, or, in other words, the reference template.
- Optical system 108' further includes a focusing lens 126' for focusing light after passage through filter 120' onto detector 110'.
- the operation of apparatus 100 is as follows. Tunable filter 120' is sequentially tuned by tuning device 122' according to a precalculated set of information, as described above, to sequentially mimic the N decorrelating matched filters, such that at selected times a different decorrelating matched filter is mimicked, while detector 110' builds an images for each until one image for each mimic has been measured. One additional image is formed while filter 120' is tuned such that light passes therethrough is undisturbed, or without filter 120' altogether.
- Equation 8 above the coordinates of the P N-dimension vector for each pixel are calculated in the N-dimensional space of the N PCs. These coordinates are then used for classification by employing the pre-prepared reference template, similarly to as described above under classification. As before, for color presentation, pixels attributed to each of the 24 different chromosome types are given a different artificial color to yield a 24 color karyotype.
- the apparatus of the present example has advantages over the apparatus of
- Example 9 in two respects.
- the apparatus according to this example has no moving parts.
- the apparatus according to this Example is less "dedicated". That is to say, should a different experimental procedure employed for chromosome painting or banding, a new set of information is calculated to permit the tunable filter to mimic a different set of decorrelation matched filters, suitable for data collection from the chromosomes according to the methods of the present invention and as detailed above.
- the operation of the apparatus of the present Example is highly suitable for computer control, which can control the operation of tuning device 122'.
- a single apparatus can be made suitable for classification and analysis of painted and banded chromosomes painted or banded by various experimental procedures, simply by employing a matching software which includes an appropriate set of information for confrolling the operation of device 122', and therefore of tunable filter 120'.
- a field of view is similar to the chromosomes sample described above. It includes scenes, each of which is spectrally different from the other just like the chromosomes when painted as described. Therefore, the principles of chromosome classification as exemplified above, apply also to remote sensing of scenes in a field of view, with the provision that each of the scenes differs from the other scenes in some specfral characteristics. This, however, is the case as described in the background section above.
- the main differences between the measurement of metaphase chromosomes and remote scenes can be summarized as follows:
- a set of N decorrelation matched filters can be calculated and manufactured for remote sensing of scenes in a field of view. These filters can be used for fast collection of decorrelated projection of
- each pixel spectrum in a field of view onto a number of orthogonal PCs 15 each pixel spectrum in a field of view onto a number of orthogonal PCs. These N values, collectively form an N-dimension vector for each pixel.
- Each of these vectors is then compared to reference vectors forming a reference template for classification, and, based on this comparison, each pixel is attributed to a specific scene, and for presentation, is given a specifying artificial color.
- a specfral decorrelation measurement apparatus or apparatus 100 For ease of measurement, the N decorrelation matched filters are placed in an apparatus referred hereinbelow as a specfral decorrelation measurement apparatus or apparatus 100".
- Apparatus 100 is connected to a telescope 101".
- a remote field of view 104" to be analyzed is viewed via telescope 101".
- Apparatus 100" further includes an optical system
- a detector 110 typically a two dimensional CCD array.
- optical system 108" may further include a collimating lens 118" to ensure full collimation of the light.
- some telescopes include a collimating lens
- collimating lens 118" may be discarded.
- Optical system 108" further includes N (N is an integer greater than two, preferably N is in the range of 3-5) decorrelating matched filters 120", five of which are shown in Figure 12, peripherally arranged on a rotatable filter carrying element 122", such as a filter wheel.
- N is an integer greater than two, preferably N is in the range of 3-5
- Rotatable filter carrying element 122" also includes one position 124" through which light passes undisturbed.
- the number N of decorrelating matched filters 120" may vary and is determined by the number of eigenvalues or PCs employed to construct the reference vector for each of the scenes in the field of view.
- Optical system 108" further includes a focusing lens 126" for focusing light after passage through rotatable filter carrying element 122" onto detector 110".
- the filters' 102" positions are preferably selected to be on the collimated section of the beam, but not necessarily: they can also be between the focusing lens and the detector (in this case, near the detector the beam size is smallest, and therefore this position is preferred).
- apparatus 100 The operation of apparatus 100" is as follows. Decorrelating matched filters 120" of rotatable filter carrying element 122" are kept successively in the radiation beam while detector 110" builds an images for each. That is to say that detector 110" builds an image with first filter 120", then rotatable filter carrying element 122” rotates to present another filter 120", and detector 110" starts building a new image in synchronization, and so on until one image for each filter 120" has been measured. One additional image is formed while position 124" through which light passes undisturbed is positioned in the path of light.
- Equation 8 the coordinates of the P N-dimension vector for each pixel are calculated in the N-dimensional space of the N PCs. These coordinates are then used for classification by employing the pre-prepared reference template. For color presentation, pixels attributed to each scene in the field of view are given a different artificial color to distinct among the scenes.
- the filter wheel may be replaced by a tunable filter and a suitable tuning device.
Abstract
Description
Claims
Priority Applications (5)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
DE69813248T DE69813248D1 (en) | 1997-07-01 | 1998-06-29 | METHOD AND DEVICE FOR REMOTE MONITORING ANALYSIS BY STATISTICAL DECORRELATION ANALYSIS |
AU81765/98A AU8176598A (en) | 1997-07-01 | 1998-06-29 | Method for remote sensing analysis by decorrelation statistical analysis and hardware therefor |
EP98931721A EP1008098B1 (en) | 1997-07-01 | 1998-06-29 | Method for remote sensing analysis by decorrelation statistical analysis and hardware therefor |
IL13350798A IL133507A0 (en) | 1997-07-01 | 1998-06-29 | Method for remote sensing analysis by decorrelation statistical analysis and hardware therefor |
JP50728299A JP2002511942A (en) | 1997-07-01 | 1998-06-29 | Remote sensing and analysis method by decorrelation statistical analysis and its hardware |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US08/886,293 US6018587A (en) | 1991-02-21 | 1997-07-01 | Method for remote sensing analysis be decorrelation statistical analysis and hardware therefor |
US08/886,293 | 1997-07-01 |
Publications (2)
Publication Number | Publication Date |
---|---|
WO1999001843A1 true WO1999001843A1 (en) | 1999-01-14 |
WO1999001843A9 WO1999001843A9 (en) | 1999-04-15 |
Family
ID=25388796
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/US1998/013524 WO1999001843A1 (en) | 1997-07-01 | 1998-06-29 | Method for remote sensing analysis by decorrelation statistical analysis and hardware therefor |
Country Status (7)
Country | Link |
---|---|
US (1) | US6018587A (en) |
EP (1) | EP1008098B1 (en) |
JP (1) | JP2002511942A (en) |
AU (1) | AU8176598A (en) |
DE (1) | DE69813248D1 (en) |
IL (1) | IL133507A0 (en) |
WO (1) | WO1999001843A1 (en) |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1296871C (en) * | 2004-08-19 | 2007-01-24 | 复旦大学 | Remote sensitive image fusing method based on residual error |
CN100380395C (en) * | 2006-06-29 | 2008-04-09 | 武汉大学 | Supervised classification process of artificial immunity in remote sensing images |
CN100416599C (en) * | 2006-06-29 | 2008-09-03 | 武汉大学 | Not supervised classification process of artificial immunity in remote sensing images |
CN100416598C (en) * | 2006-06-29 | 2008-09-03 | 武汉大学 | Method for selecting features of artificial immunity in remote sensing images |
US20080215235A1 (en) * | 2006-10-09 | 2008-09-04 | Marek Strassenburg-Kleciak | Selection and insertion of static elements in digital maps |
CN100595782C (en) * | 2008-04-17 | 2010-03-24 | 中国科学院地理科学与资源研究所 | Classification method for syncretizing optical spectrum information and multi-point simulation space information |
ITRM20090261A1 (en) * | 2009-05-22 | 2010-11-23 | Univ Calabria | AUTOMATIC METHOD OF DETECTION AND MAPPING, IN PARTICULAR OF AREAS CHARACTERIZED BY THE PRESENCE OF ASBESTOS AND / OR CEMENT-ASBESTOS, AND ITS APPARATUS. |
US9418414B2 (en) | 2012-05-30 | 2016-08-16 | Panasonic Intellectual Property Management Co., Ltd. | Image measurement apparatus, image measurement method and image measurement system |
WO2018019466A1 (en) * | 2016-07-29 | 2018-02-01 | Volkswagen Aktiengesellschaft | Method and system for detecting landmarks in a traffic environment of a mobile unit |
CN108414088A (en) * | 2017-02-10 | 2018-08-17 | 南京理工大学 | A kind of micro- polarization imaging device of EO-1 hyperion and method |
CN113807409A (en) * | 2021-08-27 | 2021-12-17 | 华北电力大学 | Coastal zone classification method based on discriminant analysis |
TWI783987B (en) * | 2017-04-14 | 2022-11-21 | 美商克萊譚克公司 | Filter apparatus, filter wheel apparatus, and inspection system |
Families Citing this family (55)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6605340B1 (en) * | 1997-02-12 | 2003-08-12 | Schweizerische Eidgenossenschaft | Camouflage structure |
CA2219809A1 (en) * | 1997-10-31 | 1999-04-30 | Shen-En Qian | System for interactive visualization and analysis of imaging spectrometry datasets over a wide-area network |
US8315909B1 (en) | 1998-03-11 | 2012-11-20 | West Corporation | Methods and apparatus for intelligent selection of goods and services in point-of-sale commerce |
US20020055906A1 (en) * | 1998-03-11 | 2002-05-09 | Katz Ronald A. | Methods and apparatus for intelligent selection of goods and services in telephonic and electronic commerce |
US7437313B1 (en) | 1998-03-11 | 2008-10-14 | West Direct, Llc | Methods, computer-readable media, and apparatus for offering users a plurality of scenarios under which to conduct at least one primary transaction |
US7364068B1 (en) | 1998-03-11 | 2008-04-29 | West Corporation | Methods and apparatus for intelligent selection of goods and services offered to conferees |
US6085227A (en) * | 1998-03-20 | 2000-07-04 | International Business Machines Corporation | System and method for operating scientific instruments over wide area networks |
US8041121B1 (en) * | 1998-08-14 | 2011-10-18 | The United States Of America As Represented By The Secretary Of The Air Force | Algorithm for rapid endmembers determination (ALRED) in hyperspectral data |
US7219086B2 (en) * | 1999-04-09 | 2007-05-15 | Plain Sight Systems, Inc. | System and method for hyper-spectral analysis |
US6738502B1 (en) | 1999-06-04 | 2004-05-18 | Kairos Scientific, Inc. | Multispectral taxonomic identification |
GB0104203D0 (en) * | 2001-02-21 | 2001-04-11 | Secr Defence | Calibrating radiometers |
AUPR631601A0 (en) * | 2001-07-11 | 2001-08-02 | Commonwealth Scientific And Industrial Research Organisation | Biotechnology array analysis |
GB0118389D0 (en) * | 2001-07-27 | 2001-09-19 | Royal Holloway And Bedford Col | Method and means for detecting internal larval infestation in granular material |
AU2002332967B2 (en) * | 2001-10-17 | 2008-07-17 | Commonwealth Scientific And Industrial Research Organisation | Method and apparatus for identifying diagnostic components of a system |
US7308139B2 (en) * | 2002-07-12 | 2007-12-11 | Chroma Energy, Inc. | Method, system, and apparatus for color representation of seismic data and associated measurements |
US7184991B1 (en) | 2002-07-12 | 2007-02-27 | Chroma Energy, Inc. | Pattern recognition applied to oil exploration and production |
US7188092B2 (en) * | 2002-07-12 | 2007-03-06 | Chroma Energy, Inc. | Pattern recognition template application applied to oil exploration and production |
US7162463B1 (en) | 2002-07-12 | 2007-01-09 | Chroma Energy, Inc. | Pattern recognition template construction applied to oil exploration and production |
US20050288863A1 (en) * | 2002-07-12 | 2005-12-29 | Chroma Energy, Inc. | Method and system for utilizing string-length ratio in seismic analysis |
US20060036410A1 (en) * | 2004-07-29 | 2006-02-16 | Ning Pan | System and method for fabric hand evaluation |
US7490009B2 (en) * | 2004-08-03 | 2009-02-10 | Fei Company | Method and system for spectroscopic data analysis |
US7631839B1 (en) * | 2004-08-20 | 2009-12-15 | Lockheed Martin Corporation | Enhanced multiple instrument distributed aperture sensor |
US7178720B1 (en) | 2004-09-30 | 2007-02-20 | West Corporation | Methods, computer-readable media, and computer program product for intelligent selection of items encoded onto portable machine-playable entertainment media |
US7186978B2 (en) * | 2004-10-15 | 2007-03-06 | Millennium Enginerring And Integration Company | Compact emissivity and temperature measuring infrared detector |
KR100561872B1 (en) * | 2004-10-22 | 2006-03-17 | 삼성전자주식회사 | Apparatus and method for analyzing component using microscopic regions |
US7689035B2 (en) * | 2005-06-17 | 2010-03-30 | The Regents Of The University Of California | Methods for identifying, separating and editing reflection components in multi-channel images and videos |
US7869043B2 (en) * | 2006-01-06 | 2011-01-11 | Sparta, Inc. | Automated passive skin detection system through spectral measurement |
US7907278B1 (en) | 2006-11-06 | 2011-03-15 | Lowell Williams | Staring imaging grating spectrometer for detection of projectiles |
CN102084229B (en) * | 2008-02-06 | 2014-10-15 | Fei公司 | A method and system for spectrum data analysis |
GB0809252D0 (en) * | 2008-05-21 | 2008-06-25 | Ntnu Technology Transfer As | Underwater hyperspectral imaging |
US20110052019A1 (en) * | 2009-09-03 | 2011-03-03 | Green Vision Systems Ltd. | Analyzing Objects Via Hyper-Spectral Imaging and Analysis |
JP5384453B2 (en) | 2010-09-09 | 2014-01-08 | シャープ株式会社 | Measuring device, measuring system, measuring method, control program, and recording medium |
US20140081605A1 (en) * | 2011-06-09 | 2014-03-20 | Kyoto University | Dtm estimation method, dtm estimation program, dtm estimation device, and method for creating 3-dimensional building model, and region extraction method, region extraction program, and region extraction device |
CN102721650B (en) * | 2012-06-13 | 2014-05-14 | 中国地质科学院矿产资源研究所 | Method and device for extracting mineral composition remote sensing information based on characteristic indexes |
US8664595B2 (en) | 2012-06-28 | 2014-03-04 | Fei Company | Cluster analysis of unknowns in SEM-EDS dataset |
US9188555B2 (en) | 2012-07-30 | 2015-11-17 | Fei Company | Automated EDS standards calibration |
US9048067B2 (en) | 2012-10-26 | 2015-06-02 | Fei Company | Mineral identification using sequential decomposition into elements from mineral definitions |
US8937282B2 (en) | 2012-10-26 | 2015-01-20 | Fei Company | Mineral identification using mineral definitions including variability |
US9778215B2 (en) | 2012-10-26 | 2017-10-03 | Fei Company | Automated mineral classification |
US9091635B2 (en) | 2012-10-26 | 2015-07-28 | Fei Company | Mineral identification using mineral definitions having compositional ranges |
US9194829B2 (en) | 2012-12-28 | 2015-11-24 | Fei Company | Process for performing automated mineralogy |
US10701287B1 (en) * | 2013-05-23 | 2020-06-30 | Rockwell Collins, Inc. | Passive clear air turbulence detection system and method |
US9714908B2 (en) | 2013-11-06 | 2017-07-25 | Fei Company | Sub-pixel analysis and display of fine grained mineral samples |
US10249214B1 (en) * | 2015-06-17 | 2019-04-02 | Vlad Joseph Novotny | Personal wellness monitoring system |
KR102508831B1 (en) * | 2016-02-17 | 2023-03-10 | 삼성전자주식회사 | Remote image transmission system, display apparatus and guide displaying method of thereof |
CN106404169B (en) * | 2016-08-26 | 2018-03-30 | 江苏天瑞仪器股份有限公司 | A kind of crystal switching device for spectrometer with indexing module |
CN106248206B (en) * | 2016-08-26 | 2017-09-12 | 江苏天瑞仪器股份有限公司 | It is a kind of to be used for the crystal switching device that spectral detection has indexing mechanism |
CN106248205B (en) * | 2016-08-26 | 2017-09-12 | 江苏天瑞仪器股份有限公司 | One kind is used for spectrometer and indexes crystal switching device |
CN106289520B (en) * | 2016-08-26 | 2017-09-29 | 江苏天瑞仪器股份有限公司 | One kind is used for spectral detection crystal switching device |
CN107515044A (en) * | 2017-09-21 | 2017-12-26 | 中国科学院云南天文台 | The atmospheric parameter monitoring device and method of the big visual field sky imaging technique of multiband |
CN108178314B (en) * | 2017-12-01 | 2021-08-24 | 东北师范大学 | Multifunctional artificial wetland experimental box and experimental method |
US11067448B2 (en) | 2018-10-05 | 2021-07-20 | Parsons Corporation | Spectral object detection |
US11099077B1 (en) * | 2019-06-04 | 2021-08-24 | The United States Of America, As Represented By The Secretary Of The Navy | Background subtracted spectrometer for airborne infrared radiometry |
CN110472357B (en) * | 2019-08-21 | 2023-05-02 | 华北理工大学 | Construction method and application method of remote sensing comprehensive ecological model RSIEI for evaluating differential effect of surface thermal environment in mining development dense area |
CN114120137B (en) * | 2021-10-19 | 2023-07-25 | 桂林理工大学 | Time-sequence vegetation remote sensing image-based wetland element time-space evolution monitoring method |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US4025772A (en) * | 1974-03-13 | 1977-05-24 | James Nickolas Constant | Digital convolver matched filter and correlator |
US4418393A (en) * | 1980-10-27 | 1983-11-29 | Sperry Corporation | Matched filter spread spectrum code recovery apparatus |
US4839843A (en) * | 1986-04-14 | 1989-06-13 | U.S. Philips Corporation | Method and apparatus for correcting a sequence of invalid samples of an equidistantly sampled signal |
US4906940A (en) * | 1987-08-24 | 1990-03-06 | Science Applications International Corporation | Process and apparatus for the automatic detection and extraction of features in images and displays |
US5287272A (en) * | 1988-04-08 | 1994-02-15 | Neuromedical Systems, Inc. | Automated cytological specimen classification system and method |
US5729471A (en) * | 1995-03-31 | 1998-03-17 | The Regents Of The University Of California | Machine dynamic selection of one video camera/image of a scene from multiple video cameras/images of the scene in accordance with a particular perspective on the scene, an object in the scene, or an event in the scene |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US4167729A (en) * | 1976-06-14 | 1979-09-11 | Environmental Research Institute Of Michigan | Apparatus for obtaining multi-spectral signatures |
US5719024A (en) * | 1993-08-18 | 1998-02-17 | Applied Spectral Imaging Ltd. | Method for chromosome classification by decorrelation statistical analysis and hardware therefore |
US5834203A (en) * | 1997-08-25 | 1998-11-10 | Applied Spectral Imaging | Method for classification of pixels into groups according to their spectra using a plurality of wide band filters and hardwire therefore |
-
1997
- 1997-07-01 US US08/886,293 patent/US6018587A/en not_active Expired - Fee Related
-
1998
- 1998-06-29 JP JP50728299A patent/JP2002511942A/en active Pending
- 1998-06-29 EP EP98931721A patent/EP1008098B1/en not_active Expired - Lifetime
- 1998-06-29 WO PCT/US1998/013524 patent/WO1999001843A1/en active IP Right Grant
- 1998-06-29 DE DE69813248T patent/DE69813248D1/en not_active Expired - Lifetime
- 1998-06-29 IL IL13350798A patent/IL133507A0/en unknown
- 1998-06-29 AU AU81765/98A patent/AU8176598A/en not_active Abandoned
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US4025772A (en) * | 1974-03-13 | 1977-05-24 | James Nickolas Constant | Digital convolver matched filter and correlator |
US4418393A (en) * | 1980-10-27 | 1983-11-29 | Sperry Corporation | Matched filter spread spectrum code recovery apparatus |
US4839843A (en) * | 1986-04-14 | 1989-06-13 | U.S. Philips Corporation | Method and apparatus for correcting a sequence of invalid samples of an equidistantly sampled signal |
US4906940A (en) * | 1987-08-24 | 1990-03-06 | Science Applications International Corporation | Process and apparatus for the automatic detection and extraction of features in images and displays |
US5287272A (en) * | 1988-04-08 | 1994-02-15 | Neuromedical Systems, Inc. | Automated cytological specimen classification system and method |
US5287272B1 (en) * | 1988-04-08 | 1996-08-27 | Neuromedical Systems Inc | Automated cytological specimen classification system and method |
US5729471A (en) * | 1995-03-31 | 1998-03-17 | The Regents Of The University Of California | Machine dynamic selection of one video camera/image of a scene from multiple video cameras/images of the scene in accordance with a particular perspective on the scene, an object in the scene, or an event in the scene |
US5745126A (en) * | 1995-03-31 | 1998-04-28 | The Regents Of The University Of California | Machine synthesis of a virtual video camera/image of a scene from multiple video cameras/images of the scene in accordance with a particular perspective on the scene, an object in the scene, or an event in the scene |
Non-Patent Citations (1)
Title |
---|
See also references of EP1008098A4 * |
Cited By (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1296871C (en) * | 2004-08-19 | 2007-01-24 | 复旦大学 | Remote sensitive image fusing method based on residual error |
CN100380395C (en) * | 2006-06-29 | 2008-04-09 | 武汉大学 | Supervised classification process of artificial immunity in remote sensing images |
CN100416599C (en) * | 2006-06-29 | 2008-09-03 | 武汉大学 | Not supervised classification process of artificial immunity in remote sensing images |
CN100416598C (en) * | 2006-06-29 | 2008-09-03 | 武汉大学 | Method for selecting features of artificial immunity in remote sensing images |
US8918274B2 (en) * | 2006-10-09 | 2014-12-23 | Harman Becker Automotive Systems Gmbh | Selection and insertion of static elements in digital maps |
US20080215235A1 (en) * | 2006-10-09 | 2008-09-04 | Marek Strassenburg-Kleciak | Selection and insertion of static elements in digital maps |
CN100595782C (en) * | 2008-04-17 | 2010-03-24 | 中国科学院地理科学与资源研究所 | Classification method for syncretizing optical spectrum information and multi-point simulation space information |
ITRM20090261A1 (en) * | 2009-05-22 | 2010-11-23 | Univ Calabria | AUTOMATIC METHOD OF DETECTION AND MAPPING, IN PARTICULAR OF AREAS CHARACTERIZED BY THE PRESENCE OF ASBESTOS AND / OR CEMENT-ASBESTOS, AND ITS APPARATUS. |
US9418414B2 (en) | 2012-05-30 | 2016-08-16 | Panasonic Intellectual Property Management Co., Ltd. | Image measurement apparatus, image measurement method and image measurement system |
WO2018019466A1 (en) * | 2016-07-29 | 2018-02-01 | Volkswagen Aktiengesellschaft | Method and system for detecting landmarks in a traffic environment of a mobile unit |
US11092444B2 (en) | 2016-07-29 | 2021-08-17 | Volkswagen Aktiengesellschaft | Method and system for recording landmarks in a traffic environment of a mobile unit |
CN108414088A (en) * | 2017-02-10 | 2018-08-17 | 南京理工大学 | A kind of micro- polarization imaging device of EO-1 hyperion and method |
TWI783987B (en) * | 2017-04-14 | 2022-11-21 | 美商克萊譚克公司 | Filter apparatus, filter wheel apparatus, and inspection system |
CN113807409A (en) * | 2021-08-27 | 2021-12-17 | 华北电力大学 | Coastal zone classification method based on discriminant analysis |
CN113807409B (en) * | 2021-08-27 | 2024-01-23 | 华北电力大学 | Coastal zone classification method based on discriminant analysis |
Also Published As
Publication number | Publication date |
---|---|
IL133507A0 (en) | 2001-04-30 |
AU8176598A (en) | 1999-01-25 |
JP2002511942A (en) | 2002-04-16 |
EP1008098A1 (en) | 2000-06-14 |
EP1008098B1 (en) | 2003-04-09 |
EP1008098A4 (en) | 2000-11-22 |
WO1999001843A9 (en) | 1999-04-15 |
US6018587A (en) | 2000-01-25 |
DE69813248D1 (en) | 2003-05-15 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
EP1008098B1 (en) | Method for remote sensing analysis by decorrelation statistical analysis and hardware therefor | |
Jia et al. | Status and application of advanced airborne hyperspectral imaging technology: A review | |
Zarco-Tejada et al. | Chlorophyll fluorescence effects on vegetation apparent reflectance: II. Laboratory and airborne canopy-level measurements with hyperspectral data | |
US5991028A (en) | Spectral bio-imaging methods for cell classification | |
Sòria et al. | ENVISAT/AATSR derived land surface temperature over a heterogeneous region | |
Gomez | Hyperspectral imaging: a useful technology for transportation analysis | |
Reddy | Satellite remote sensing sensors: principles and applications | |
Colwell | Some practical applications of multiband spectral reconnaissance | |
Barducci et al. | Aerospace wetland monitoring by hyperspectral imaging sensors: A case study in the coastal zone of San Rossore Natural Park | |
Zhu et al. | SPAD monitoring of saline vegetation based on Gaussian mixture model and UAV hyperspectral image feature classification | |
Navalgund | Remote sensing: 1. Basics and applications | |
Curtis | Remote sensing systems for monitoring crops and vegetation | |
Giardino et al. | Airborne remote sensing and geospatial analysis | |
Li et al. | Predicting water content using linear spectral mixture model on soil spectra | |
Roderick et al. | Remote sensing in vegetation and animal studies. | |
Zhang | Surface water quality estimation using remote sensing in the Gulf of Finland and the Finnish Archipelago Sea | |
Suomalainen | Empirical Studies on Multiangular, Hyperspectral, and Polarimetric Reflectance of Natural Surfaces | |
Gritskevich et al. | The Use of Digital Cameras for Multispectral Registration with an Unmanned Aircraft | |
Morain et al. | Fundamentals of Electromagnetic Radiation | |
Kattel | Study of Seasonal Change and Water Stress Condition in Plant Leaf Using Polarimetric Lidar Measurement | |
Gupta et al. | Interpretation of Solar Reflection Data | |
Jakubowski et al. | Synthetic data generation of high-resolution hyperspectral data using DIRSIG | |
Melillos et al. | Detecting Underground Military Structures Using Field Spectroscopy | |
Jones | Wavelet image processing applied to hyperspectral images for foliage penetration | |
Baumgardner et al. | Remote sensing techniques used in agriculture today |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
WWE | Wipo information: entry into national phase |
Ref document number: 133507 Country of ref document: IL |
|
AK | Designated states |
Kind code of ref document: A1 Designated state(s): AL AM AT AU AZ BA BB BG BR BY CA CH CN CU CZ DE DK EE ES FI GB GE GH GM GW HU ID IL IS JP KE KG KP KR KZ LC LK LR LS LT LU LV MD MG MK MN MW MX NO NZ PL PT RO RU SD SE SG SI SK SL TJ TM TR TT UA UG US UZ VN YU ZW |
|
AL | Designated countries for regional patents |
Kind code of ref document: A1 Designated state(s): GH GM KE LS MW SD SZ UG ZW AM AZ BY KG KZ MD RU TJ TM AT BE CH CY DE DK ES FI FR GB GR IE IT LU MC NL PT SE BF BJ CF CG CI CM GA GN ML MR NE SN TD TG |
|
AK | Designated states |
Kind code of ref document: C2 Designated state(s): AL AM AT AU AZ BA BB BG BR BY CA CH CN CU CZ DE DK EE ES FI GB GE GH GM GW HU ID IL IS JP KE KG KP KR KZ LC LK LR LS LT LU LV MD MG MK MN MW MX NO NZ PL PT RO RU SD SE SG SI SK SL TJ TM TR TT UA UG US UZ VN YU ZW |
|
AL | Designated countries for regional patents |
Kind code of ref document: C2 Designated state(s): GH GM KE LS MW SD SZ UG ZW AM AZ BY KG KZ MD RU TJ TM AT BE CH CY DE DK ES FI FR GB GR IE IT LU MC NL PT SE BF BJ CF CG CI CM GA GN ML MR NE SN TD TG |
|
COP | Corrected version of pamphlet |
Free format text: PAGES 1/16-16/16, DRAWINGS, REPLACED BY NEW PAGES 1/16-16/16; DUE TO LATE TRANSMITTAL BY THE RECEIVING OFFICE |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application | ||
DFPE | Request for preliminary examination filed prior to expiration of 19th month from priority date (pct application filed before 20040101) | ||
NENP | Non-entry into the national phase |
Ref country code: KR |
|
WWE | Wipo information: entry into national phase |
Ref document number: 1998931721 Country of ref document: EP |
|
REG | Reference to national code |
Ref country code: DE Ref legal event code: 8642 |
|
WWP | Wipo information: published in national office |
Ref document number: 1998931721 Country of ref document: EP |
|
NENP | Non-entry into the national phase |
Ref country code: CA |
|
WWG | Wipo information: grant in national office |
Ref document number: 1998931721 Country of ref document: EP |